Upload folder using huggingface_hub

.gitattributes

@@ -1285,3 +1285,6 @@ classic_pythia-2.8b-step2/imdb:model=EleutherAI_pythia-2.8b,data_augmentation=ca
 classic_pythia-2.8b-step2/imdb:model=EleutherAI_pythia-2.8b,data_augmentation=canonical/per_instance_stats.json filter=lfs diff=lfs merge=lfs -text
 classic_pythia-2.8b-step2/imdb:model=EleutherAI_pythia-2.8b,data_augmentation=canonical/scenario_state.json filter=lfs diff=lfs merge=lfs -text
 classic_pythia-2.8b-step2/runs.json filter=lfs diff=lfs merge=lfs -text
+mmlu_pythia-1.4b-step2000/mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law/display_requests.json filter=lfs diff=lfs merge=lfs -text
+mmlu_pythia-1.4b-step2000/mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law/per_instance_stats.json filter=lfs diff=lfs merge=lfs -text
+mmlu_pythia-1.4b-step2000/mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law/scenario_state.json filter=lfs diff=lfs merge=lfs -text

mmlu_pythia-1.4b-step2000/costs.json

@@ -0,0 +1 @@
+{}

mmlu_pythia-1.4b-step2000/groups.json

@@ -0,0 +1,2411 @@
+[
+  {
+    "title": "All Scenarios",
+    "header": [
+      {
+        "value": "Group",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "Description",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "Adaptation method",
+        "description": "Adaptation strategy (e.g., generation)",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# instances",
+        "description": "Number of instances evaluated on",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# references",
+        "description": "Number of references provided per instance",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# prompt tokens",
+        "description": "Total number of prompt tokens",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# completion tokens",
+        "description": "Total number of completion tokens",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# models",
+        "description": "Number of models we're evaluating",
+        "markdown": false,
+        "metadata": {}
+      }
+    ],
+    "rows": [
+      [
+        {
+          "value": "MMLU Subjects",
+          "href": "?group=mmlu_subjects",
+          "markdown": false
+        },
+        {
+          "value": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 246.35087719298247,
+          "description": "min=100, mean=246.351, max=1534, sum=28084 (114)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=1368 (342)",
+          "markdown": false
+        },
+        {
+          "value": 204487.2446201024,
+          "description": "min=285.766, mean=597.916, max=1816.758, sum=204487.245 (342)",
+          "markdown": false
+        },
+        {
+          "value": 342.0,
+          "description": "min=1, mean=1, max=1, sum=342 (342)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ]
+    ],
+    "links": []
+  },
+  {
+    "title": "Scenarios",
+    "header": [
+      {
+        "value": "Group",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "Description",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "Adaptation method",
+        "description": "Adaptation strategy (e.g., generation)",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# instances",
+        "description": "Number of instances evaluated on",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# references",
+        "description": "Number of references provided per instance",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# prompt tokens",
+        "description": "Total number of prompt tokens",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# completion tokens",
+        "description": "Total number of completion tokens",
+        "markdown": false,
+        "metadata": {}
+      },
+      {
+        "value": "# models",
+        "description": "Number of models we're evaluating",
+        "markdown": false,
+        "metadata": {}
+      }
+    ],
+    "rows": [
+      [
+        {
+          "value": "Massive Multitask Language Understanding (MMLU) All Subjects",
+          "href": "?group=mmlu",
+          "markdown": false
+        },
+        {
+          "value": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 246.35087719298247,
+          "description": "min=100, mean=246.351, max=1534, sum=14042 (57)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=684 (171)",
+          "markdown": false
+        },
+        {
+          "value": 102243.62231005121,
+          "description": "min=285.766, mean=597.916, max=1816.758, sum=102243.622 (171)",
+          "markdown": false
+        },
+        {
+          "value": 171.0,
+          "description": "min=1, mean=1, max=1, sum=171 (171)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Abstract Algebra",
+          "href": "?group=mmlu_abstract_algebra",
+          "markdown": false
+        },
+        {
+          "value": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1076.28,
+          "description": "min=358.76, mean=358.76, max=358.76, sum=1076.28 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Anatomy",
+          "href": "?group=mmlu_anatomy",
+          "markdown": false
+        },
+        {
+          "value": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 135.0,
+          "description": "min=135, mean=135, max=135, sum=135 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1056.088888888889,
+          "description": "min=352.03, mean=352.03, max=352.03, sum=1056.089 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "College Chemistry",
+          "href": "?group=mmlu_college_chemistry",
+          "markdown": false
+        },
+        {
+          "value": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1607.5500000000002,
+          "description": "min=535.85, mean=535.85, max=535.85, sum=1607.55 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Computer Security",
+          "href": "?group=mmlu_computer_security",
+          "markdown": false
+        },
+        {
+          "value": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1164.57,
+          "description": "min=388.19, mean=388.19, max=388.19, sum=1164.57 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Econometrics",
+          "href": "?group=mmlu_econometrics",
+          "markdown": false
+        },
+        {
+          "value": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 114.0,
+          "description": "min=114, mean=114, max=114, sum=114 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1838.3947368421052,
+          "description": "min=612.798, mean=612.798, max=612.798, sum=1838.395 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Global Facts",
+          "href": "?group=mmlu_global_facts",
+          "markdown": false
+        },
+        {
+          "value": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1201.74,
+          "description": "min=400.58, mean=400.58, max=400.58, sum=1201.74 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Jurisprudence",
+          "href": "?group=mmlu_jurisprudence",
+          "markdown": false
+        },
+        {
+          "value": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 108.0,
+          "description": "min=108, mean=108, max=108, sum=108 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1262.5833333333333,
+          "description": "min=420.861, mean=420.861, max=420.861, sum=1262.583 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Philosophy",
+          "href": "?group=mmlu_philosophy",
+          "markdown": false
+        },
+        {
+          "value": "The philosophy subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 311.0,
+          "description": "min=311, mean=311, max=311, sum=311 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1035.829581993569,
+          "description": "min=345.277, mean=345.277, max=345.277, sum=1035.83 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Professional Medicine",
+          "href": "?group=mmlu_professional_medicine",
+          "markdown": false
+        },
+        {
+          "value": "The professional medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 272.0,
+          "description": "min=272, mean=272, max=272, sum=272 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3242.6470588235297,
+          "description": "min=1080.882, mean=1080.882, max=1080.882, sum=3242.647 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Us Foreign Policy",
+          "href": "?group=mmlu_us_foreign_policy",
+          "markdown": false
+        },
+        {
+          "value": "The us foreign policy subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1332.24,
+          "description": "min=444.08, mean=444.08, max=444.08, sum=1332.24 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Astronomy",
+          "href": "?group=mmlu_astronomy",
+          "markdown": false
+        },
+        {
+          "value": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 152.0,
+          "description": "min=152, mean=152, max=152, sum=152 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1795.4605263157894,
+          "description": "min=598.487, mean=598.487, max=598.487, sum=1795.461 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Business Ethics",
+          "href": "?group=mmlu_business_ethics",
+          "markdown": false
+        },
+        {
+          "value": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1755.1499999999999,
+          "description": "min=585.05, mean=585.05, max=585.05, sum=1755.15 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Clinical Knowledge",
+          "href": "?group=mmlu_clinical_knowledge",
+          "markdown": false
+        },
+        {
+          "value": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 265.0,
+          "description": "min=265, mean=265, max=265, sum=265 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1205.7509433962264,
+          "description": "min=401.917, mean=401.917, max=401.917, sum=1205.751 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "College Biology",
+          "href": "?group=mmlu_college_biology",
+          "markdown": false
+        },
+        {
+          "value": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 144.0,
+          "description": "min=144, mean=144, max=144, sum=144 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1410.9583333333335,
+          "description": "min=470.319, mean=470.319, max=470.319, sum=1410.958 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "College Computer Science",
+          "href": "?group=mmlu_college_computer_science",
+          "markdown": false
+        },
+        {
+          "value": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 2528.67,
+          "description": "min=842.89, mean=842.89, max=842.89, sum=2528.67 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "College Mathematics",
+          "href": "?group=mmlu_college_mathematics",
+          "markdown": false
+        },
+        {
+          "value": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1778.46,
+          "description": "min=592.82, mean=592.82, max=592.82, sum=1778.46 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "College Medicine",
+          "href": "?group=mmlu_college_medicine",
+          "markdown": false
+        },
+        {
+          "value": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 173.0,
+          "description": "min=173, mean=173, max=173, sum=173 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1558.1271676300578,
+          "description": "min=519.376, mean=519.376, max=519.376, sum=1558.127 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "College Physics",
+          "href": "?group=mmlu_college_physics",
+          "markdown": false
+        },
+        {
+          "value": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 102.0,
+          "description": "min=102, mean=102, max=102, sum=102 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1429.9705882352941,
+          "description": "min=476.657, mean=476.657, max=476.657, sum=1429.971 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Conceptual Physics",
+          "href": "?group=mmlu_conceptual_physics",
+          "markdown": false
+        },
+        {
+          "value": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 235.0,
+          "description": "min=235, mean=235, max=235, sum=235 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 933.931914893617,
+          "description": "min=311.311, mean=311.311, max=311.311, sum=933.932 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Electrical Engineering",
+          "href": "?group=mmlu_electrical_engineering",
+          "markdown": false
+        },
+        {
+          "value": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 145.0,
+          "description": "min=145, mean=145, max=145, sum=145 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1274.544827586207,
+          "description": "min=424.848, mean=424.848, max=424.848, sum=1274.545 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Elementary Mathematics",
+          "href": "?group=mmlu_elementary_mathematics",
+          "markdown": false
+        },
+        {
+          "value": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 378.0,
+          "description": "min=378, mean=378, max=378, sum=378 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1515.2142857142858,
+          "description": "min=505.071, mean=505.071, max=505.071, sum=1515.214 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Formal Logic",
+          "href": "?group=mmlu_formal_logic",
+          "markdown": false
+        },
+        {
+          "value": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 126.0,
+          "description": "min=126, mean=126, max=126, sum=126 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1960.7857142857142,
+          "description": "min=653.595, mean=653.595, max=653.595, sum=1960.786 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Biology",
+          "href": "?group=mmlu_high_school_biology",
+          "markdown": false
+        },
+        {
+          "value": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 310.0,
+          "description": "min=310, mean=310, max=310, sum=310 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1541.7967741935481,
+          "description": "min=513.932, mean=513.932, max=513.932, sum=1541.797 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Chemistry",
+          "href": "?group=mmlu_high_school_chemistry",
+          "markdown": false
+        },
+        {
+          "value": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 203.0,
+          "description": "min=203, mean=203, max=203, sum=203 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1439.527093596059,
+          "description": "min=479.842, mean=479.842, max=479.842, sum=1439.527 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Computer Science",
+          "href": "?group=mmlu_high_school_computer_science",
+          "markdown": false
+        },
+        {
+          "value": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 2668.17,
+          "description": "min=889.39, mean=889.39, max=889.39, sum=2668.17 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School European History",
+          "href": "?group=mmlu_high_school_european_history",
+          "markdown": false
+        },
+        {
+          "value": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 165.0,
+          "description": "min=165, mean=165, max=165, sum=165 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 5450.272727272727,
+          "description": "min=1816.758, mean=1816.758, max=1816.758, sum=5450.273 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Geography",
+          "href": "?group=mmlu_high_school_geography",
+          "markdown": false
+        },
+        {
+          "value": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 198.0,
+          "description": "min=198, mean=198, max=198, sum=198 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1200.2727272727273,
+          "description": "min=400.091, mean=400.091, max=400.091, sum=1200.273 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Government And Politics",
+          "href": "?group=mmlu_high_school_government_and_politics",
+          "markdown": false
+        },
+        {
+          "value": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 193.0,
+          "description": "min=193, mean=193, max=193, sum=193 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1448.2849740932643,
+          "description": "min=482.762, mean=482.762, max=482.762, sum=1448.285 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Macroeconomics",
+          "href": "?group=mmlu_high_school_macroeconomics",
+          "markdown": false
+        },
+        {
+          "value": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 390.0,
+          "description": "min=390, mean=390, max=390, sum=390 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1177.0538461538463,
+          "description": "min=392.351, mean=392.351, max=392.351, sum=1177.054 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Mathematics",
+          "href": "?group=mmlu_high_school_mathematics",
+          "markdown": false
+        },
+        {
+          "value": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 270.0,
+          "description": "min=270, mean=270, max=270, sum=270 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1520.0666666666666,
+          "description": "min=506.689, mean=506.689, max=506.689, sum=1520.067 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Microeconomics",
+          "href": "?group=mmlu_high_school_microeconomics",
+          "markdown": false
+        },
+        {
+          "value": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 238.0,
+          "description": "min=238, mean=238, max=238, sum=238 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1233.7058823529412,
+          "description": "min=411.235, mean=411.235, max=411.235, sum=1233.706 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Physics",
+          "href": "?group=mmlu_high_school_physics",
+          "markdown": false
+        },
+        {
+          "value": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 151.0,
+          "description": "min=151, mean=151, max=151, sum=151 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1646.1854304635763,
+          "description": "min=548.728, mean=548.728, max=548.728, sum=1646.185 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Psychology",
+          "href": "?group=mmlu_high_school_psychology",
+          "markdown": false
+        },
+        {
+          "value": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 545.0,
+          "description": "min=545, mean=545, max=545, sum=545 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1544.377981651376,
+          "description": "min=514.793, mean=514.793, max=514.793, sum=1544.378 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School Statistics",
+          "href": "?group=mmlu_high_school_statistics",
+          "markdown": false
+        },
+        {
+          "value": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 216.0,
+          "description": "min=216, mean=216, max=216, sum=216 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 2389.8194444444443,
+          "description": "min=796.606, mean=796.606, max=796.606, sum=2389.819 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School US History",
+          "href": "?group=mmlu_high_school_us_history",
+          "markdown": false
+        },
+        {
+          "value": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 204.0,
+          "description": "min=204, mean=204, max=204, sum=204 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 5365.161764705883,
+          "description": "min=1788.387, mean=1788.387, max=1788.387, sum=5365.162 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "High School World History",
+          "href": "?group=mmlu_high_school_world_history",
+          "markdown": false
+        },
+        {
+          "value": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 237.0,
+          "description": "min=237, mean=237, max=237, sum=237 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 4384.329113924051,
+          "description": "min=1461.443, mean=1461.443, max=1461.443, sum=4384.329 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Human Aging",
+          "href": "?group=mmlu_human_aging",
+          "markdown": false
+        },
+        {
+          "value": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 223.0,
+          "description": "min=223, mean=223, max=223, sum=223 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 971.0717488789237,
+          "description": "min=323.691, mean=323.691, max=323.691, sum=971.072 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Human Sexuality",
+          "href": "?group=mmlu_human_sexuality",
+          "markdown": false
+        },
+        {
+          "value": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 131.0,
+          "description": "min=131, mean=131, max=131, sum=131 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1066.0534351145038,
+          "description": "min=355.351, mean=355.351, max=355.351, sum=1066.053 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "International Law",
+          "href": "?group=mmlu_international_law",
+          "markdown": false
+        },
+        {
+          "value": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 121.0,
+          "description": "min=121, mean=121, max=121, sum=121 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1951.115702479339,
+          "description": "min=650.372, mean=650.372, max=650.372, sum=1951.116 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Logical Fallacies",
+          "href": "?group=mmlu_logical_fallacies",
+          "markdown": false
+        },
+        {
+          "value": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 163.0,
+          "description": "min=163, mean=163, max=163, sum=163 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1376.484662576687,
+          "description": "min=458.828, mean=458.828, max=458.828, sum=1376.485 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Machine Learning",
+          "href": "?group=mmlu_machine_learning",
+          "markdown": false
+        },
+        {
+          "value": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 112.0,
+          "description": "min=112, mean=112, max=112, sum=112 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1983.6428571428569,
+          "description": "min=661.214, mean=661.214, max=661.214, sum=1983.643 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Management",
+          "href": "?group=mmlu_management",
+          "markdown": false
+        },
+        {
+          "value": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 103.0,
+          "description": "min=103, mean=103, max=103, sum=103 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 894.1456310679612,
+          "description": "min=298.049, mean=298.049, max=298.049, sum=894.146 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Marketing",
+          "href": "?group=mmlu_marketing",
+          "markdown": false
+        },
+        {
+          "value": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 234.0,
+          "description": "min=234, mean=234, max=234, sum=234 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1320.3076923076924,
+          "description": "min=440.103, mean=440.103, max=440.103, sum=1320.308 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Medical Genetics",
+          "href": "?group=mmlu_medical_genetics",
+          "markdown": false
+        },
+        {
+          "value": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 100.0,
+          "description": "min=100, mean=100, max=100, sum=100 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1021.44,
+          "description": "min=340.48, mean=340.48, max=340.48, sum=1021.44 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Miscellaneous",
+          "href": "?group=mmlu_miscellaneous",
+          "markdown": false
+        },
+        {
+          "value": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 783.0,
+          "description": "min=783, mean=783, max=783, sum=783 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 961.3295019157088,
+          "description": "min=320.443, mean=320.443, max=320.443, sum=961.33 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Moral Disputes",
+          "href": "?group=mmlu_moral_disputes",
+          "markdown": false
+        },
+        {
+          "value": "The moral disputes subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 346.0,
+          "description": "min=346, mean=346, max=346, sum=346 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1506.728323699422,
+          "description": "min=502.243, mean=502.243, max=502.243, sum=1506.728 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Moral Scenarios",
+          "href": "?group=mmlu_moral_scenarios",
+          "markdown": false
+        },
+        {
+          "value": "The moral scenarios subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 895.0,
+          "description": "min=895, mean=895, max=895, sum=895 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 2003.5843575418994,
+          "description": "min=667.861, mean=667.861, max=667.861, sum=2003.584 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Nutrition",
+          "href": "?group=mmlu_nutrition",
+          "markdown": false
+        },
+        {
+          "value": "The nutrition subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 306.0,
+          "description": "min=306, mean=306, max=306, sum=306 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1737.3823529411766,
+          "description": "min=579.127, mean=579.127, max=579.127, sum=1737.382 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Prehistory",
+          "href": "?group=mmlu_prehistory",
+          "markdown": false
+        },
+        {
+          "value": "The prehistory subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 324.0,
+          "description": "min=324, mean=324, max=324, sum=324 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1605.453703703704,
+          "description": "min=535.151, mean=535.151, max=535.151, sum=1605.454 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Professional Accounting",
+          "href": "?group=mmlu_professional_accounting",
+          "markdown": false
+        },
+        {
+          "value": "The professional accounting subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 282.0,
+          "description": "min=282, mean=282, max=282, sum=282 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1982.7659574468084,
+          "description": "min=660.922, mean=660.922, max=660.922, sum=1982.766 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Professional Law",
+          "href": "?group=mmlu_professional_law",
+          "markdown": false
+        },
+        {
+          "value": "The professional law subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 1534.0,
+          "description": "min=1534, mean=1534, max=1534, sum=1534 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 4963.298565840939,
+          "description": "min=1654.433, mean=1654.433, max=1654.433, sum=4963.299 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Professional Psychology",
+          "href": "?group=mmlu_professional_psychology",
+          "markdown": false
+        },
+        {
+          "value": "The professional psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 612.0,
+          "description": "min=612, mean=612, max=612, sum=612 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1772.6176470588234,
+          "description": "min=590.873, mean=590.873, max=590.873, sum=1772.618 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Public Relations",
+          "href": "?group=mmlu_public_relations",
+          "markdown": false
+        },
+        {
+          "value": "The public relations subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 110.0,
+          "description": "min=110, mean=110, max=110, sum=110 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1268.9454545454546,
+          "description": "min=422.982, mean=422.982, max=422.982, sum=1268.945 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Security Studies",
+          "href": "?group=mmlu_security_studies",
+          "markdown": false
+        },
+        {
+          "value": "The security studies subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 245.0,
+          "description": "min=245, mean=245, max=245, sum=245 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3621.171428571429,
+          "description": "min=1207.057, mean=1207.057, max=1207.057, sum=3621.171 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Sociology",
+          "href": "?group=mmlu_sociology",
+          "markdown": false
+        },
+        {
+          "value": "The sociology subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 201.0,
+          "description": "min=201, mean=201, max=201, sum=201 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1356.0597014925372,
+          "description": "min=452.02, mean=452.02, max=452.02, sum=1356.06 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "Virology",
+          "href": "?group=mmlu_virology",
+          "markdown": false
+        },
+        {
+          "value": "The virology subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 166.0,
+          "description": "min=166, mean=166, max=166, sum=166 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1048.7530120481927,
+          "description": "min=349.584, mean=349.584, max=349.584, sum=1048.753 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ],
+      [
+        {
+          "value": "World Religions",
+          "href": "?group=mmlu_world_religions",
+          "markdown": false
+        },
+        {
+          "value": "The world religions subject in the Massive Multitask Language Understanding (MMLU) benchmark.",
+          "markdown": true
+        },
+        {
+          "value": "multiple_choice_joint",
+          "markdown": false
+        },
+        {
+          "value": 171.0,
+          "description": "min=171, mean=171, max=171, sum=171 (1)",
+          "markdown": false
+        },
+        {
+          "value": 4.0,
+          "description": "min=4, mean=4, max=4, sum=12 (3)",
+          "markdown": false
+        },
+        {
+          "value": 857.2982456140351,
+          "description": "min=285.766, mean=285.766, max=285.766, sum=857.298 (3)",
+          "markdown": false
+        },
+        {
+          "value": 3.0,
+          "description": "min=1, mean=1, max=1, sum=3 (3)",
+          "markdown": false
+        },
+        {
+          "value": 1,
+          "markdown": false
+        }
+      ]
+    ],
+    "links": []
+  }
+]

mmlu_pythia-1.4b-step2000/groups/json/mmlu_abstract_algebra_mmlu_abstract_algebra_subject:abstract_algebra.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: abstract_algebra",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Abstract Algebra"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Abstract Algebra"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Abstract Algebra"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Abstract Algebra"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Abstract Algebra"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Abstract Algebra"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The abstract algebra subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Abstract Algebra"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_abstract_algebra&subgroup=subject%3A%20abstract_algebra&runSpecs=%5B%22mmlu%3Asubject%3Dabstract_algebra%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_abstract_algebra%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra"
+        ]
+      },
+      {
+        "value": 0.23,
+        "description": "min=0.23, mean=0.23, max=0.23, sum=0.23 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06573188066482544,
+        "description": "min=0.066, mean=0.066, max=0.066, sum=0.066 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 358.76,
+        "description": "min=358.76, mean=358.76, max=358.76, sum=358.76 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_abstract_algebra_mmlu_abstract_algebra_subject:abstract_algebra.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_abstract_algebra_mmlu_abstract_algebra_subject:abstract_algebra.json"
+    }
+  ],
+  "name": "mmlu_abstract_algebra_subject:abstract_algebra"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_anatomy_mmlu_anatomy_subject:anatomy.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: anatomy",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Anatomy"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Anatomy"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Anatomy"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Anatomy"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Anatomy"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Anatomy"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The anatomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Anatomy"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_anatomy&subgroup=subject%3A%20anatomy&runSpecs=%5B%22mmlu%3Asubject%3Danatomy%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_anatomy%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy"
+        ]
+      },
+      {
+        "value": 0.2962962962962963,
+        "description": "min=0.296, mean=0.296, max=0.296, sum=0.296 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06830824922632288,
+        "description": "min=0.068, mean=0.068, max=0.068, sum=0.068 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 135.0,
+        "description": "min=135, mean=135, max=135, sum=135 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 352.02962962962965,
+        "description": "min=352.03, mean=352.03, max=352.03, sum=352.03 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_anatomy_mmlu_anatomy_subject:anatomy.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_anatomy_mmlu_anatomy_subject:anatomy.json"
+    }
+  ],
+  "name": "mmlu_anatomy_subject:anatomy"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_astronomy_mmlu_astronomy_subject:astronomy.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: astronomy",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Astronomy"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Astronomy"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Astronomy"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Astronomy"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Astronomy"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Astronomy"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The astronomy subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Astronomy"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_astronomy&subgroup=subject%3A%20astronomy&runSpecs=%5B%22mmlu%3Asubject%3Dastronomy%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_astronomy%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy"
+        ]
+      },
+      {
+        "value": 0.23684210526315788,
+        "description": "min=0.237, mean=0.237, max=0.237, sum=0.237 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.10126333958224247,
+        "description": "min=0.101, mean=0.101, max=0.101, sum=0.101 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 152.0,
+        "description": "min=152, mean=152, max=152, sum=152 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 598.4868421052631,
+        "description": "min=598.487, mean=598.487, max=598.487, sum=598.487 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_astronomy_mmlu_astronomy_subject:astronomy.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_astronomy_mmlu_astronomy_subject:astronomy.json"
+    }
+  ],
+  "name": "mmlu_astronomy_subject:astronomy"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_business_ethics_mmlu_business_ethics_subject:business_ethics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: business_ethics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Business Ethics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Business Ethics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Business Ethics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Business Ethics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Business Ethics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Business Ethics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The business ethics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Business Ethics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_business_ethics&subgroup=subject%3A%20business_ethics&runSpecs=%5B%22mmlu%3Asubject%3Dbusiness_ethics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_business_ethics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics"
+        ]
+      },
+      {
+        "value": 0.29,
+        "description": "min=0.29, mean=0.29, max=0.29, sum=0.29 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.09688833236694336,
+        "description": "min=0.097, mean=0.097, max=0.097, sum=0.097 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 585.05,
+        "description": "min=585.05, mean=585.05, max=585.05, sum=585.05 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_business_ethics_mmlu_business_ethics_subject:business_ethics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_business_ethics_mmlu_business_ethics_subject:business_ethics.json"
+    }
+  ],
+  "name": "mmlu_business_ethics_subject:business_ethics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_clinical_knowledge_mmlu_clinical_knowledge_subject:clinical_knowledge.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: clinical_knowledge",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Clinical Knowledge"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Clinical Knowledge"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Clinical Knowledge"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Clinical Knowledge"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Clinical Knowledge"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Clinical Knowledge"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The clinical knowledge subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Clinical Knowledge"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_clinical_knowledge&subgroup=subject%3A%20clinical_knowledge&runSpecs=%5B%22mmlu%3Asubject%3Dclinical_knowledge%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_clinical_knowledge%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge"
+        ]
+      },
+      {
+        "value": 0.2037735849056604,
+        "description": "min=0.204, mean=0.204, max=0.204, sum=0.204 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.0775772769496126,
+        "description": "min=0.078, mean=0.078, max=0.078, sum=0.078 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 265.0,
+        "description": "min=265, mean=265, max=265, sum=265 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 401.9169811320755,
+        "description": "min=401.917, mean=401.917, max=401.917, sum=401.917 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_clinical_knowledge_mmlu_clinical_knowledge_subject:clinical_knowledge.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_clinical_knowledge_mmlu_clinical_knowledge_subject:clinical_knowledge.json"
+    }
+  ],
+  "name": "mmlu_clinical_knowledge_subject:clinical_knowledge"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_biology_mmlu_college_biology_subject:college_biology.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: college_biology",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "College Biology"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "College Biology"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "College Biology"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "College Biology"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "College Biology"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "College Biology"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The college biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "College Biology"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_college_biology&subgroup=subject%3A%20college_biology&runSpecs=%5B%22mmlu%3Asubject%3Dcollege_biology%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_college_biology%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology"
+        ]
+      },
+      {
+        "value": 0.3333333333333333,
+        "description": "min=0.333, mean=0.333, max=0.333, sum=0.333 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08110955854256947,
+        "description": "min=0.081, mean=0.081, max=0.081, sum=0.081 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 144.0,
+        "description": "min=144, mean=144, max=144, sum=144 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 470.31944444444446,
+        "description": "min=470.319, mean=470.319, max=470.319, sum=470.319 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_college_biology_mmlu_college_biology_subject:college_biology.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_biology_mmlu_college_biology_subject:college_biology.json"
+    }
+  ],
+  "name": "mmlu_college_biology_subject:college_biology"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_chemistry_mmlu_college_chemistry_subject:college_chemistry.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: college_chemistry",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "College Chemistry"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "College Chemistry"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "College Chemistry"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "College Chemistry"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "College Chemistry"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "College Chemistry"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The college chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "College Chemistry"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_college_chemistry&subgroup=subject%3A%20college_chemistry&runSpecs=%5B%22mmlu%3Asubject%3Dcollege_chemistry%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_college_chemistry%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry"
+        ]
+      },
+      {
+        "value": 0.26,
+        "description": "min=0.26, mean=0.26, max=0.26, sum=0.26 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08982811689376831,
+        "description": "min=0.09, mean=0.09, max=0.09, sum=0.09 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 535.85,
+        "description": "min=535.85, mean=535.85, max=535.85, sum=535.85 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_college_chemistry_mmlu_college_chemistry_subject:college_chemistry.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_chemistry_mmlu_college_chemistry_subject:college_chemistry.json"
+    }
+  ],
+  "name": "mmlu_college_chemistry_subject:college_chemistry"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_computer_science_mmlu_college_computer_science_subject:college_computer_science.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: college_computer_science",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "College Computer Science"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "College Computer Science"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "College Computer Science"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "College Computer Science"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "College Computer Science"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "College Computer Science"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The college computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "College Computer Science"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_college_computer_science&subgroup=subject%3A%20college_computer_science&runSpecs=%5B%22mmlu%3Asubject%3Dcollege_computer_science%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_college_computer_science%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science"
+        ]
+      },
+      {
+        "value": 0.35,
+        "description": "min=0.35, mean=0.35, max=0.35, sum=0.35 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.13751029014587401,
+        "description": "min=0.138, mean=0.138, max=0.138, sum=0.138 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 842.89,
+        "description": "min=842.89, mean=842.89, max=842.89, sum=842.89 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_college_computer_science_mmlu_college_computer_science_subject:college_computer_science.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_computer_science_mmlu_college_computer_science_subject:college_computer_science.json"
+    }
+  ],
+  "name": "mmlu_college_computer_science_subject:college_computer_science"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_mathematics_mmlu_college_mathematics_subject:college_mathematics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: college_mathematics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "College Mathematics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "College Mathematics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "College Mathematics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "College Mathematics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "College Mathematics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "College Mathematics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The college mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "College Mathematics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_college_mathematics&subgroup=subject%3A%20college_mathematics&runSpecs=%5B%22mmlu%3Asubject%3Dcollege_mathematics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_college_mathematics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics"
+        ]
+      },
+      {
+        "value": 0.2,
+        "description": "min=0.2, mean=0.2, max=0.2, sum=0.2 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.10322931528091431,
+        "description": "min=0.103, mean=0.103, max=0.103, sum=0.103 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 592.82,
+        "description": "min=592.82, mean=592.82, max=592.82, sum=592.82 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_college_mathematics_mmlu_college_mathematics_subject:college_mathematics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_mathematics_mmlu_college_mathematics_subject:college_mathematics.json"
+    }
+  ],
+  "name": "mmlu_college_mathematics_subject:college_mathematics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_medicine_mmlu_college_medicine_subject:college_medicine.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: college_medicine",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "College Medicine"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "College Medicine"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "College Medicine"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "College Medicine"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "College Medicine"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "College Medicine"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The college medicine subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "College Medicine"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_college_medicine&subgroup=subject%3A%20college_medicine&runSpecs=%5B%22mmlu%3Asubject%3Dcollege_medicine%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_college_medicine%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine"
+        ]
+      },
+      {
+        "value": 0.23699421965317918,
+        "description": "min=0.237, mean=0.237, max=0.237, sum=0.237 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08715281045505766,
+        "description": "min=0.087, mean=0.087, max=0.087, sum=0.087 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 173.0,
+        "description": "min=173, mean=173, max=173, sum=173 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 519.3757225433526,
+        "description": "min=519.376, mean=519.376, max=519.376, sum=519.376 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_college_medicine_mmlu_college_medicine_subject:college_medicine.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_medicine_mmlu_college_medicine_subject:college_medicine.json"
+    }
+  ],
+  "name": "mmlu_college_medicine_subject:college_medicine"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_physics_mmlu_college_physics_subject:college_physics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: college_physics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "College Physics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "College Physics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "College Physics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "College Physics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "College Physics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "College Physics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The college physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "College Physics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_college_physics&subgroup=subject%3A%20college_physics&runSpecs=%5B%22mmlu%3Asubject%3Dcollege_physics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_college_physics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics"
+        ]
+      },
+      {
+        "value": 0.22549019607843138,
+        "description": "min=0.225, mean=0.225, max=0.225, sum=0.225 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08214688534830131,
+        "description": "min=0.082, mean=0.082, max=0.082, sum=0.082 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 102.0,
+        "description": "min=102, mean=102, max=102, sum=102 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 476.65686274509807,
+        "description": "min=476.657, mean=476.657, max=476.657, sum=476.657 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_college_physics_mmlu_college_physics_subject:college_physics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_college_physics_mmlu_college_physics_subject:college_physics.json"
+    }
+  ],
+  "name": "mmlu_college_physics_subject:college_physics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_computer_security_mmlu_computer_security_subject:computer_security.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: computer_security",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Computer Security"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Computer Security"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Computer Security"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Computer Security"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Computer Security"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Computer Security"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The computer security subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Computer Security"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_computer_security&subgroup=subject%3A%20computer_security&runSpecs=%5B%22mmlu%3Asubject%3Dcomputer_security%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_computer_security%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security"
+        ]
+      },
+      {
+        "value": 0.22,
+        "description": "min=0.22, mean=0.22, max=0.22, sum=0.22 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07067869663238525,
+        "description": "min=0.071, mean=0.071, max=0.071, sum=0.071 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 388.19,
+        "description": "min=388.19, mean=388.19, max=388.19, sum=388.19 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_computer_security_mmlu_computer_security_subject:computer_security.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_computer_security_mmlu_computer_security_subject:computer_security.json"
+    }
+  ],
+  "name": "mmlu_computer_security_subject:computer_security"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_conceptual_physics_mmlu_conceptual_physics_subject:conceptual_physics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: conceptual_physics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Conceptual Physics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Conceptual Physics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Conceptual Physics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Conceptual Physics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Conceptual Physics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Conceptual Physics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The conceptual physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Conceptual Physics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_conceptual_physics&subgroup=subject%3A%20conceptual_physics&runSpecs=%5B%22mmlu%3Asubject%3Dconceptual_physics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_conceptual_physics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics"
+        ]
+      },
+      {
+        "value": 0.2170212765957447,
+        "description": "min=0.217, mean=0.217, max=0.217, sum=0.217 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06520833461842639,
+        "description": "min=0.065, mean=0.065, max=0.065, sum=0.065 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 235.0,
+        "description": "min=235, mean=235, max=235, sum=235 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 311.31063829787234,
+        "description": "min=311.311, mean=311.311, max=311.311, sum=311.311 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_conceptual_physics_mmlu_conceptual_physics_subject:conceptual_physics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_conceptual_physics_mmlu_conceptual_physics_subject:conceptual_physics.json"
+    }
+  ],
+  "name": "mmlu_conceptual_physics_subject:conceptual_physics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_econometrics_mmlu_econometrics_subject:econometrics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: econometrics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Econometrics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Econometrics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Econometrics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Econometrics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Econometrics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Econometrics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The econometrics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Econometrics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_econometrics&subgroup=subject%3A%20econometrics&runSpecs=%5B%22mmlu%3Asubject%3Deconometrics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_econometrics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics"
+        ]
+      },
+      {
+        "value": 0.24561403508771928,
+        "description": "min=0.246, mean=0.246, max=0.246, sum=0.246 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.1076611740547314,
+        "description": "min=0.108, mean=0.108, max=0.108, sum=0.108 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 114.0,
+        "description": "min=114, mean=114, max=114, sum=114 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 612.7982456140351,
+        "description": "min=612.798, mean=612.798, max=612.798, sum=612.798 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_econometrics_mmlu_econometrics_subject:econometrics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_econometrics_mmlu_econometrics_subject:econometrics.json"
+    }
+  ],
+  "name": "mmlu_econometrics_subject:econometrics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_electrical_engineering_mmlu_electrical_engineering_subject:electrical_engineering.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: electrical_engineering",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Electrical Engineering"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Electrical Engineering"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Electrical Engineering"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Electrical Engineering"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Electrical Engineering"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Electrical Engineering"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The electrical engineering subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Electrical Engineering"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_electrical_engineering&subgroup=subject%3A%20electrical_engineering&runSpecs=%5B%22mmlu%3Asubject%3Delectrical_engineering%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_electrical_engineering%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering"
+        ]
+      },
+      {
+        "value": 0.2827586206896552,
+        "description": "min=0.283, mean=0.283, max=0.283, sum=0.283 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07903624896345468,
+        "description": "min=0.079, mean=0.079, max=0.079, sum=0.079 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 145.0,
+        "description": "min=145, mean=145, max=145, sum=145 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 424.848275862069,
+        "description": "min=424.848, mean=424.848, max=424.848, sum=424.848 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_electrical_engineering_mmlu_electrical_engineering_subject:electrical_engineering.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_electrical_engineering_mmlu_electrical_engineering_subject:electrical_engineering.json"
+    }
+  ],
+  "name": "mmlu_electrical_engineering_subject:electrical_engineering"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_elementary_mathematics_mmlu_elementary_mathematics_subject:elementary_mathematics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: elementary_mathematics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Elementary Mathematics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Elementary Mathematics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Elementary Mathematics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Elementary Mathematics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Elementary Mathematics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Elementary Mathematics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The elementary mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Elementary Mathematics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_elementary_mathematics&subgroup=subject%3A%20elementary_mathematics&runSpecs=%5B%22mmlu%3Asubject%3Delementary_mathematics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_elementary_mathematics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics"
+        ]
+      },
+      {
+        "value": 0.2328042328042328,
+        "description": "min=0.233, mean=0.233, max=0.233, sum=0.233 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08565072975461445,
+        "description": "min=0.086, mean=0.086, max=0.086, sum=0.086 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 378.0,
+        "description": "min=378, mean=378, max=378, sum=378 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 505.07142857142856,
+        "description": "min=505.071, mean=505.071, max=505.071, sum=505.071 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_elementary_mathematics_mmlu_elementary_mathematics_subject:elementary_mathematics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_elementary_mathematics_mmlu_elementary_mathematics_subject:elementary_mathematics.json"
+    }
+  ],
+  "name": "mmlu_elementary_mathematics_subject:elementary_mathematics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_formal_logic_mmlu_formal_logic_subject:formal_logic.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: formal_logic",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Formal Logic"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Formal Logic"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Formal Logic"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Formal Logic"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Formal Logic"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Formal Logic"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The formal logic subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Formal Logic"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_formal_logic&subgroup=subject%3A%20formal_logic&runSpecs=%5B%22mmlu%3Asubject%3Dformal_logic%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_formal_logic%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic"
+        ]
+      },
+      {
+        "value": 0.23015873015873015,
+        "description": "min=0.23, mean=0.23, max=0.23, sum=0.23 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.11266928627377465,
+        "description": "min=0.113, mean=0.113, max=0.113, sum=0.113 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 126.0,
+        "description": "min=126, mean=126, max=126, sum=126 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 653.5952380952381,
+        "description": "min=653.595, mean=653.595, max=653.595, sum=653.595 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_formal_logic_mmlu_formal_logic_subject:formal_logic.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_formal_logic_mmlu_formal_logic_subject:formal_logic.json"
+    }
+  ],
+  "name": "mmlu_formal_logic_subject:formal_logic"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_global_facts_mmlu_global_facts_subject:global_facts.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: global_facts",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Global Facts"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Global Facts"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Global Facts"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Global Facts"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Global Facts"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Global Facts"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The global facts subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Global Facts"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_global_facts&subgroup=subject%3A%20global_facts&runSpecs=%5B%22mmlu%3Asubject%3Dglobal_facts%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_global_facts%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts"
+        ]
+      },
+      {
+        "value": 0.28,
+        "description": "min=0.28, mean=0.28, max=0.28, sum=0.28 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07499703168869018,
+        "description": "min=0.075, mean=0.075, max=0.075, sum=0.075 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 400.58,
+        "description": "min=400.58, mean=400.58, max=400.58, sum=400.58 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_global_facts_mmlu_global_facts_subject:global_facts.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_global_facts_mmlu_global_facts_subject:global_facts.json"
+    }
+  ],
+  "name": "mmlu_global_facts_subject:global_facts"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_biology_mmlu_high_school_biology_subject:high_school_biology.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_biology",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Biology"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Biology"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Biology"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Biology"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Biology"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Biology"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school biology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Biology"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_biology&subgroup=subject%3A%20high_school_biology&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_biology%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_biology%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology"
+        ]
+      },
+      {
+        "value": 0.21935483870967742,
+        "description": "min=0.219, mean=0.219, max=0.219, sum=0.219 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08429911059717979,
+        "description": "min=0.084, mean=0.084, max=0.084, sum=0.084 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 310.0,
+        "description": "min=310, mean=310, max=310, sum=310 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 513.9322580645161,
+        "description": "min=513.932, mean=513.932, max=513.932, sum=513.932 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_biology_mmlu_high_school_biology_subject:high_school_biology.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_biology_mmlu_high_school_biology_subject:high_school_biology.json"
+    }
+  ],
+  "name": "mmlu_high_school_biology_subject:high_school_biology"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_chemistry_mmlu_high_school_chemistry_subject:high_school_chemistry.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_chemistry",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Chemistry"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Chemistry"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Chemistry"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Chemistry"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Chemistry"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Chemistry"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school chemistry subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Chemistry"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_chemistry&subgroup=subject%3A%20high_school_chemistry&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_chemistry%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_chemistry%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry"
+        ]
+      },
+      {
+        "value": 0.27586206896551724,
+        "description": "min=0.276, mean=0.276, max=0.276, sum=0.276 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08463767714101106,
+        "description": "min=0.085, mean=0.085, max=0.085, sum=0.085 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 203.0,
+        "description": "min=203, mean=203, max=203, sum=203 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 479.8423645320197,
+        "description": "min=479.842, mean=479.842, max=479.842, sum=479.842 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_chemistry_mmlu_high_school_chemistry_subject:high_school_chemistry.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_chemistry_mmlu_high_school_chemistry_subject:high_school_chemistry.json"
+    }
+  ],
+  "name": "mmlu_high_school_chemistry_subject:high_school_chemistry"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_computer_science_mmlu_high_school_computer_science_subject:high_school_computer_science.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_computer_science",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Computer Science"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Computer Science"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Computer Science"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Computer Science"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Computer Science"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Computer Science"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school computer science subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Computer Science"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_computer_science&subgroup=subject%3A%20high_school_computer_science&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_computer_science%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_computer_science%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science"
+        ]
+      },
+      {
+        "value": 0.17,
+        "description": "min=0.17, mean=0.17, max=0.17, sum=0.17 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.14540541887283326,
+        "description": "min=0.145, mean=0.145, max=0.145, sum=0.145 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 889.39,
+        "description": "min=889.39, mean=889.39, max=889.39, sum=889.39 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_computer_science_mmlu_high_school_computer_science_subject:high_school_computer_science.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_computer_science_mmlu_high_school_computer_science_subject:high_school_computer_science.json"
+    }
+  ],
+  "name": "mmlu_high_school_computer_science_subject:high_school_computer_science"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_european_history_mmlu_high_school_european_history_subject:high_school_european_history.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_european_history",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School European History"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School European History"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School European History"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School European History"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School European History"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School European History"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school european history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School European History"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_european_history&subgroup=subject%3A%20high_school_european_history&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_european_history%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_european_history%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history"
+        ]
+      },
+      {
+        "value": 0.20606060606060606,
+        "description": "min=0.206, mean=0.206, max=0.206, sum=0.206 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.3232730460889412,
+        "description": "min=0.323, mean=0.323, max=0.323, sum=0.323 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 165.0,
+        "description": "min=165, mean=165, max=165, sum=165 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 2.909090909090909,
+        "description": "min=2.909, mean=2.909, max=2.909, sum=2.909 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1816.7575757575758,
+        "description": "min=1816.758, mean=1816.758, max=1816.758, sum=1816.758 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_european_history_mmlu_high_school_european_history_subject:high_school_european_history.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_european_history_mmlu_high_school_european_history_subject:high_school_european_history.json"
+    }
+  ],
+  "name": "mmlu_high_school_european_history_subject:high_school_european_history"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_geography_mmlu_high_school_geography_subject:high_school_geography.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_geography",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Geography"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Geography"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Geography"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Geography"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Geography"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Geography"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school geography subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Geography"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_geography&subgroup=subject%3A%20high_school_geography&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_geography%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_geography%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography"
+        ]
+      },
+      {
+        "value": 0.2828282828282828,
+        "description": "min=0.283, mean=0.283, max=0.283, sum=0.283 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07503375019690003,
+        "description": "min=0.075, mean=0.075, max=0.075, sum=0.075 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 198.0,
+        "description": "min=198, mean=198, max=198, sum=198 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 400.09090909090907,
+        "description": "min=400.091, mean=400.091, max=400.091, sum=400.091 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_geography_mmlu_high_school_geography_subject:high_school_geography.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_geography_mmlu_high_school_geography_subject:high_school_geography.json"
+    }
+  ],
+  "name": "mmlu_high_school_geography_subject:high_school_geography"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_government_and_politics_mmlu_high_school_government_and_politics_subject:high_school_government_and_politics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_government_and_politics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Government And Politics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Government And Politics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Government And Politics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Government And Politics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Government And Politics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Government And Politics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school government and politics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Government And Politics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_government_and_politics&subgroup=subject%3A%20high_school_government_and_politics&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_government_and_politics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_government_and_politics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics"
+        ]
+      },
+      {
+        "value": 0.2849740932642487,
+        "description": "min=0.285, mean=0.285, max=0.285, sum=0.285 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08498252858769709,
+        "description": "min=0.085, mean=0.085, max=0.085, sum=0.085 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 193.0,
+        "description": "min=193, mean=193, max=193, sum=193 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 482.7616580310881,
+        "description": "min=482.762, mean=482.762, max=482.762, sum=482.762 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_government_and_politics_mmlu_high_school_government_and_politics_subject:high_school_government_and_politics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_government_and_politics_mmlu_high_school_government_and_politics_subject:high_school_government_and_politics.json"
+    }
+  ],
+  "name": "mmlu_high_school_government_and_politics_subject:high_school_government_and_politics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_macroeconomics_mmlu_high_school_macroeconomics_subject:high_school_macroeconomics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_macroeconomics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Macroeconomics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Macroeconomics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Macroeconomics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Macroeconomics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Macroeconomics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Macroeconomics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school macroeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Macroeconomics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_macroeconomics&subgroup=subject%3A%20high_school_macroeconomics&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_macroeconomics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_macroeconomics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics"
+        ]
+      },
+      {
+        "value": 0.3384615384615385,
+        "description": "min=0.338, mean=0.338, max=0.338, sum=0.338 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07281008500319261,
+        "description": "min=0.073, mean=0.073, max=0.073, sum=0.073 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 390.0,
+        "description": "min=390, mean=390, max=390, sum=390 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 392.35128205128206,
+        "description": "min=392.351, mean=392.351, max=392.351, sum=392.351 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_macroeconomics_mmlu_high_school_macroeconomics_subject:high_school_macroeconomics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_macroeconomics_mmlu_high_school_macroeconomics_subject:high_school_macroeconomics.json"
+    }
+  ],
+  "name": "mmlu_high_school_macroeconomics_subject:high_school_macroeconomics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_mathematics_mmlu_high_school_mathematics_subject:high_school_mathematics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_mathematics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Mathematics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Mathematics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Mathematics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Mathematics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Mathematics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Mathematics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school mathematics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Mathematics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_mathematics&subgroup=subject%3A%20high_school_mathematics&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_mathematics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_mathematics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics"
+        ]
+      },
+      {
+        "value": 0.2518518518518518,
+        "description": "min=0.252, mean=0.252, max=0.252, sum=0.252 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08458723138879846,
+        "description": "min=0.085, mean=0.085, max=0.085, sum=0.085 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 270.0,
+        "description": "min=270, mean=270, max=270, sum=270 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 506.68888888888887,
+        "description": "min=506.689, mean=506.689, max=506.689, sum=506.689 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_mathematics_mmlu_high_school_mathematics_subject:high_school_mathematics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_mathematics_mmlu_high_school_mathematics_subject:high_school_mathematics.json"
+    }
+  ],
+  "name": "mmlu_high_school_mathematics_subject:high_school_mathematics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_microeconomics_mmlu_high_school_microeconomics_subject:high_school_microeconomics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_microeconomics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Microeconomics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Microeconomics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Microeconomics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Microeconomics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Microeconomics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Microeconomics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school microeconomics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Microeconomics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_microeconomics&subgroup=subject%3A%20high_school_microeconomics&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_microeconomics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_microeconomics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics"
+        ]
+      },
+      {
+        "value": 0.3445378151260504,
+        "description": "min=0.345, mean=0.345, max=0.345, sum=0.345 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07571322076460894,
+        "description": "min=0.076, mean=0.076, max=0.076, sum=0.076 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 238.0,
+        "description": "min=238, mean=238, max=238, sum=238 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 411.2352941176471,
+        "description": "min=411.235, mean=411.235, max=411.235, sum=411.235 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_microeconomics_mmlu_high_school_microeconomics_subject:high_school_microeconomics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_microeconomics_mmlu_high_school_microeconomics_subject:high_school_microeconomics.json"
+    }
+  ],
+  "name": "mmlu_high_school_microeconomics_subject:high_school_microeconomics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_physics_mmlu_high_school_physics_subject:high_school_physics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_physics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Physics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Physics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Physics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Physics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Physics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Physics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school physics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Physics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_physics&subgroup=subject%3A%20high_school_physics&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_physics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_physics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics"
+        ]
+      },
+      {
+        "value": 0.31125827814569534,
+        "description": "min=0.311, mean=0.311, max=0.311, sum=0.311 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.09131200739879482,
+        "description": "min=0.091, mean=0.091, max=0.091, sum=0.091 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 151.0,
+        "description": "min=151, mean=151, max=151, sum=151 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 548.7284768211921,
+        "description": "min=548.728, mean=548.728, max=548.728, sum=548.728 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_physics_mmlu_high_school_physics_subject:high_school_physics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_physics_mmlu_high_school_physics_subject:high_school_physics.json"
+    }
+  ],
+  "name": "mmlu_high_school_physics_subject:high_school_physics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_psychology_mmlu_high_school_psychology_subject:high_school_psychology.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_psychology",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Psychology"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Psychology"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Psychology"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Psychology"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Psychology"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Psychology"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school psychology subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Psychology"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_psychology&subgroup=subject%3A%20high_school_psychology&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_psychology%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_psychology%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology"
+        ]
+      },
+      {
+        "value": 0.26238532110091745,
+        "description": "min=0.262, mean=0.262, max=0.262, sum=0.262 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08488267058626227,
+        "description": "min=0.085, mean=0.085, max=0.085, sum=0.085 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 545.0,
+        "description": "min=545, mean=545, max=545, sum=545 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 514.7926605504587,
+        "description": "min=514.793, mean=514.793, max=514.793, sum=514.793 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_psychology_mmlu_high_school_psychology_subject:high_school_psychology.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_psychology_mmlu_high_school_psychology_subject:high_school_psychology.json"
+    }
+  ],
+  "name": "mmlu_high_school_psychology_subject:high_school_psychology"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_statistics_mmlu_high_school_statistics_subject:high_school_statistics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_statistics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School Statistics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School Statistics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School Statistics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School Statistics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School Statistics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School Statistics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school statistics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School Statistics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_statistics&subgroup=subject%3A%20high_school_statistics&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_statistics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_statistics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics"
+        ]
+      },
+      {
+        "value": 0.4351851851851852,
+        "description": "min=0.435, mean=0.435, max=0.435, sum=0.435 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.13173586792416042,
+        "description": "min=0.132, mean=0.132, max=0.132, sum=0.132 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 216.0,
+        "description": "min=216, mean=216, max=216, sum=216 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 796.6064814814815,
+        "description": "min=796.606, mean=796.606, max=796.606, sum=796.606 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_statistics_mmlu_high_school_statistics_subject:high_school_statistics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_statistics_mmlu_high_school_statistics_subject:high_school_statistics.json"
+    }
+  ],
+  "name": "mmlu_high_school_statistics_subject:high_school_statistics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_us_history_mmlu_high_school_us_history_subject:high_school_us_history.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_us_history",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School US History"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School US History"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School US History"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School US History"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School US History"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School US History"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school us history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School US History"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_us_history&subgroup=subject%3A%20high_school_us_history&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_us_history%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_us_history%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history"
+        ]
+      },
+      {
+        "value": 0.23529411764705882,
+        "description": "min=0.235, mean=0.235, max=0.235, sum=0.235 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.3188782766753552,
+        "description": "min=0.319, mean=0.319, max=0.319, sum=0.319 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 204.0,
+        "description": "min=204, mean=204, max=204, sum=204 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 4.0,
+        "description": "min=4, mean=4, max=4, sum=4 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1788.387254901961,
+        "description": "min=1788.387, mean=1788.387, max=1788.387, sum=1788.387 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_us_history_mmlu_high_school_us_history_subject:high_school_us_history.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_us_history_mmlu_high_school_us_history_subject:high_school_us_history.json"
+    }
+  ],
+  "name": "mmlu_high_school_us_history_subject:high_school_us_history"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_world_history_mmlu_high_school_world_history_subject:high_school_world_history.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: high_school_world_history",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "High School World History"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "High School World History"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "High School World History"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "High School World History"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "High School World History"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "High School World History"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The high school world history subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "High School World History"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_high_school_world_history&subgroup=subject%3A%20high_school_world_history&runSpecs=%5B%22mmlu%3Asubject%3Dhigh_school_world_history%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_high_school_world_history%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history"
+        ]
+      },
+      {
+        "value": 0.18143459915611815,
+        "description": "min=0.181, mean=0.181, max=0.181, sum=0.181 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.2615184371481465,
+        "description": "min=0.262, mean=0.262, max=0.262, sum=0.262 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 237.0,
+        "description": "min=237, mean=237, max=237, sum=237 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1461.4430379746836,
+        "description": "min=1461.443, mean=1461.443, max=1461.443, sum=1461.443 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_high_school_world_history_mmlu_high_school_world_history_subject:high_school_world_history.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_high_school_world_history_mmlu_high_school_world_history_subject:high_school_world_history.json"
+    }
+  ],
+  "name": "mmlu_high_school_world_history_subject:high_school_world_history"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_human_aging_mmlu_human_aging_subject:human_aging.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: human_aging",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Human Aging"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Human Aging"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Human Aging"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Human Aging"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Human Aging"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Human Aging"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The human aging subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Human Aging"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_human_aging&subgroup=subject%3A%20human_aging&runSpecs=%5B%22mmlu%3Asubject%3Dhuman_aging%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_human_aging%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging"
+        ]
+      },
+      {
+        "value": 0.16143497757847533,
+        "description": "min=0.161, mean=0.161, max=0.161, sum=0.161 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06565283766776457,
+        "description": "min=0.066, mean=0.066, max=0.066, sum=0.066 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 223.0,
+        "description": "min=223, mean=223, max=223, sum=223 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 323.69058295964123,
+        "description": "min=323.691, mean=323.691, max=323.691, sum=323.691 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_human_aging_mmlu_human_aging_subject:human_aging.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_human_aging_mmlu_human_aging_subject:human_aging.json"
+    }
+  ],
+  "name": "mmlu_human_aging_subject:human_aging"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_human_sexuality_mmlu_human_sexuality_subject:human_sexuality.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: human_sexuality",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Human Sexuality"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Human Sexuality"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Human Sexuality"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Human Sexuality"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Human Sexuality"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Human Sexuality"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The human sexuality subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Human Sexuality"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_human_sexuality&subgroup=subject%3A%20human_sexuality&runSpecs=%5B%22mmlu%3Asubject%3Dhuman_sexuality%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_human_sexuality%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality"
+        ]
+      },
+      {
+        "value": 0.2595419847328244,
+        "description": "min=0.26, mean=0.26, max=0.26, sum=0.26 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06619657268960967,
+        "description": "min=0.066, mean=0.066, max=0.066, sum=0.066 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 131.0,
+        "description": "min=131, mean=131, max=131, sum=131 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 355.35114503816794,
+        "description": "min=355.351, mean=355.351, max=355.351, sum=355.351 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_human_sexuality_mmlu_human_sexuality_subject:human_sexuality.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_human_sexuality_mmlu_human_sexuality_subject:human_sexuality.json"
+    }
+  ],
+  "name": "mmlu_human_sexuality_subject:human_sexuality"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_international_law_mmlu_international_law_subject:international_law.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: international_law",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "International Law"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "International Law"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "International Law"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "International Law"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "International Law"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "International Law"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The international law subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "International Law"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_international_law&subgroup=subject%3A%20international_law&runSpecs=%5B%22mmlu%3Asubject%3Dinternational_law%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_international_law%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law"
+        ]
+      },
+      {
+        "value": 0.256198347107438,
+        "description": "min=0.256, mean=0.256, max=0.256, sum=0.256 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.10936137073296161,
+        "description": "min=0.109, mean=0.109, max=0.109, sum=0.109 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 121.0,
+        "description": "min=121, mean=121, max=121, sum=121 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 650.3719008264463,
+        "description": "min=650.372, mean=650.372, max=650.372, sum=650.372 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_international_law_mmlu_international_law_subject:international_law.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_international_law_mmlu_international_law_subject:international_law.json"
+    }
+  ],
+  "name": "mmlu_international_law_subject:international_law"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_jurisprudence_mmlu_jurisprudence_subject:jurisprudence.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: jurisprudence",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Jurisprudence"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Jurisprudence"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Jurisprudence"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Jurisprudence"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Jurisprudence"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Jurisprudence"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The jurisprudence subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Jurisprudence"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_jurisprudence&subgroup=subject%3A%20jurisprudence&runSpecs=%5B%22mmlu%3Asubject%3Djurisprudence%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_jurisprudence%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence"
+        ]
+      },
+      {
+        "value": 0.21296296296296297,
+        "description": "min=0.213, mean=0.213, max=0.213, sum=0.213 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.07681376183474506,
+        "description": "min=0.077, mean=0.077, max=0.077, sum=0.077 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 108.0,
+        "description": "min=108, mean=108, max=108, sum=108 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 420.8611111111111,
+        "description": "min=420.861, mean=420.861, max=420.861, sum=420.861 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_jurisprudence_mmlu_jurisprudence_subject:jurisprudence.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_jurisprudence_mmlu_jurisprudence_subject:jurisprudence.json"
+    }
+  ],
+  "name": "mmlu_jurisprudence_subject:jurisprudence"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_logical_fallacies_mmlu_logical_fallacies_subject:logical_fallacies.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: logical_fallacies",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Logical Fallacies"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Logical Fallacies"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Logical Fallacies"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Logical Fallacies"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Logical Fallacies"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Logical Fallacies"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The logical fallacies subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Logical Fallacies"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_logical_fallacies&subgroup=subject%3A%20logical_fallacies&runSpecs=%5B%22mmlu%3Asubject%3Dlogical_fallacies%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_logical_fallacies%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies"
+        ]
+      },
+      {
+        "value": 0.22699386503067484,
+        "description": "min=0.227, mean=0.227, max=0.227, sum=0.227 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08394306715280732,
+        "description": "min=0.084, mean=0.084, max=0.084, sum=0.084 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 163.0,
+        "description": "min=163, mean=163, max=163, sum=163 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 458.8282208588957,
+        "description": "min=458.828, mean=458.828, max=458.828, sum=458.828 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_logical_fallacies_mmlu_logical_fallacies_subject:logical_fallacies.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_logical_fallacies_mmlu_logical_fallacies_subject:logical_fallacies.json"
+    }
+  ],
+  "name": "mmlu_logical_fallacies_subject:logical_fallacies"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_machine_learning_mmlu_machine_learning_subject:machine_learning.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: machine_learning",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Machine Learning"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Machine Learning"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Machine Learning"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Machine Learning"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Machine Learning"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Machine Learning"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The machine learning subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Machine Learning"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_machine_learning&subgroup=subject%3A%20machine_learning&runSpecs=%5B%22mmlu%3Asubject%3Dmachine_learning%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_machine_learning%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning"
+        ]
+      },
+      {
+        "value": 0.21428571428571427,
+        "description": "min=0.214, mean=0.214, max=0.214, sum=0.214 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.11298963214669909,
+        "description": "min=0.113, mean=0.113, max=0.113, sum=0.113 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 112.0,
+        "description": "min=112, mean=112, max=112, sum=112 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 661.2142857142857,
+        "description": "min=661.214, mean=661.214, max=661.214, sum=661.214 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_machine_learning_mmlu_machine_learning_subject:machine_learning.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_machine_learning_mmlu_machine_learning_subject:machine_learning.json"
+    }
+  ],
+  "name": "mmlu_machine_learning_subject:machine_learning"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_management_mmlu_management_subject:management.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: management",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Management"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Management"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Management"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Management"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Management"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Management"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The management subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Management"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_management&subgroup=subject%3A%20management&runSpecs=%5B%22mmlu%3Asubject%3Dmanagement%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_management%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management"
+        ]
+      },
+      {
+        "value": 0.17475728155339806,
+        "description": "min=0.175, mean=0.175, max=0.175, sum=0.175 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06428557923696573,
+        "description": "min=0.064, mean=0.064, max=0.064, sum=0.064 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 103.0,
+        "description": "min=103, mean=103, max=103, sum=103 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 298.0485436893204,
+        "description": "min=298.049, mean=298.049, max=298.049, sum=298.049 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_management_mmlu_management_subject:management.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_management_mmlu_management_subject:management.json"
+    }
+  ],
+  "name": "mmlu_management_subject:management"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_marketing_mmlu_marketing_subject:marketing.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: marketing",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Marketing"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Marketing"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Marketing"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Marketing"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Marketing"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Marketing"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The marketing subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Marketing"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_marketing&subgroup=subject%3A%20marketing&runSpecs=%5B%22mmlu%3Asubject%3Dmarketing%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_marketing%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing"
+        ]
+      },
+      {
+        "value": 0.23931623931623933,
+        "description": "min=0.239, mean=0.239, max=0.239, sum=0.239 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.08168192704518636,
+        "description": "min=0.082, mean=0.082, max=0.082, sum=0.082 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 234.0,
+        "description": "min=234, mean=234, max=234, sum=234 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 440.1025641025641,
+        "description": "min=440.103, mean=440.103, max=440.103, sum=440.103 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_marketing_mmlu_marketing_subject:marketing.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_marketing_mmlu_marketing_subject:marketing.json"
+    }
+  ],
+  "name": "mmlu_marketing_subject:marketing"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_medical_genetics_mmlu_medical_genetics_subject:medical_genetics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: medical_genetics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Medical Genetics"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Medical Genetics"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Medical Genetics"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Medical Genetics"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Medical Genetics"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Medical Genetics"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The medical genetics subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Medical Genetics"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_medical_genetics&subgroup=subject%3A%20medical_genetics&runSpecs=%5B%22mmlu%3Asubject%3Dmedical_genetics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_medical_genetics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics"
+        ]
+      },
+      {
+        "value": 0.31,
+        "description": "min=0.31, mean=0.31, max=0.31, sum=0.31 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06659618139266968,
+        "description": "min=0.067, mean=0.067, max=0.067, sum=0.067 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 340.48,
+        "description": "min=340.48, mean=340.48, max=340.48, sum=340.48 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_medical_genetics_mmlu_medical_genetics_subject:medical_genetics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_medical_genetics_mmlu_medical_genetics_subject:medical_genetics.json"
+    }
+  ],
+  "name": "mmlu_medical_genetics_subject:medical_genetics"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_miscellaneous_mmlu_miscellaneous_subject:miscellaneous.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: miscellaneous",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "Miscellaneous"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "Miscellaneous"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "Miscellaneous"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "Miscellaneous"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "Miscellaneous"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "Miscellaneous"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The miscellaneous subject in the Massive Multitask Language Understanding (MMLU) benchmark.\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "Miscellaneous"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu_miscellaneous&subgroup=subject%3A%20miscellaneous&runSpecs=%5B%22mmlu%3Asubject%3Dmiscellaneous%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_miscellaneous%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous"
+        ]
+      },
+      {
+        "value": 0.26436781609195403,
+        "description": "min=0.264, mean=0.264, max=0.264, sum=0.264 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06466218734000682,
+        "description": "min=0.065, mean=0.065, max=0.065, sum=0.065 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 783.0,
+        "description": "min=783, mean=783, max=783, sum=783 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 320.4431673052363,
+        "description": "min=320.443, mean=320.443, max=320.443, sum=320.443 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_miscellaneous_mmlu_miscellaneous_subject:miscellaneous.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_miscellaneous_mmlu_miscellaneous_subject:miscellaneous.json"
+    }
+  ],
+  "name": "mmlu_miscellaneous_subject:miscellaneous"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu.json

@@ -0,0 +1,554 @@
+{
+  "title": "Massive Multitask Language Understanding (MMLU) All Subjects",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "markdown": false
+      },
+      {
+        "value": 0.2514631519798282,
+        "description": "min=0.161, mean=0.251, max=0.438, sum=14.333 (57)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      },
+      {
+        "value": 0.1056762656473459,
+        "description": "min=0.061, mean=0.106, max=0.323, sum=6.024 (57)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      },
+      {
+        "value": 246.35087719298247,
+        "description": "min=100, mean=246.351, max=1534, sum=14042 (57)",
+        "style": {},
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      },
+      {
+        "value": 4.945727778020373,
+        "description": "min=2.909, mean=4.946, max=5, sum=281.906 (57)",
+        "style": {},
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (57)",
+        "style": {},
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      },
+      {
+        "value": 597.9159199418197,
+        "description": "min=285.766, mean=597.916, max=1816.758, sum=34081.207 (57)",
+        "style": {},
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=57 (57)",
+        "style": {},
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra",
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy",
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy",
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics",
+          "mmlu:subject=clinical_knowledge,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_clinical_knowledge",
+          "mmlu:subject=college_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_biology",
+          "mmlu:subject=college_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_chemistry",
+          "mmlu:subject=college_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_computer_science",
+          "mmlu:subject=college_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_mathematics",
+          "mmlu:subject=college_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_medicine",
+          "mmlu:subject=college_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_college_physics",
+          "mmlu:subject=computer_security,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_computer_security",
+          "mmlu:subject=conceptual_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_conceptual_physics",
+          "mmlu:subject=econometrics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_econometrics",
+          "mmlu:subject=electrical_engineering,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_electrical_engineering",
+          "mmlu:subject=elementary_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_elementary_mathematics",
+          "mmlu:subject=formal_logic,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_formal_logic",
+          "mmlu:subject=global_facts,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_global_facts",
+          "mmlu:subject=high_school_biology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_biology",
+          "mmlu:subject=high_school_chemistry,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_chemistry",
+          "mmlu:subject=high_school_computer_science,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_computer_science",
+          "mmlu:subject=high_school_european_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_european_history",
+          "mmlu:subject=high_school_geography,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_geography",
+          "mmlu:subject=high_school_government_and_politics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_government_and_politics",
+          "mmlu:subject=high_school_macroeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_macroeconomics",
+          "mmlu:subject=high_school_mathematics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_mathematics",
+          "mmlu:subject=high_school_microeconomics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_microeconomics",
+          "mmlu:subject=high_school_physics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_physics",
+          "mmlu:subject=high_school_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_psychology",
+          "mmlu:subject=high_school_statistics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_statistics",
+          "mmlu:subject=high_school_us_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_us_history",
+          "mmlu:subject=high_school_world_history,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_high_school_world_history",
+          "mmlu:subject=human_aging,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_aging",
+          "mmlu:subject=human_sexuality,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_human_sexuality",
+          "mmlu:subject=international_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_international_law",
+          "mmlu:subject=jurisprudence,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_jurisprudence",
+          "mmlu:subject=logical_fallacies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_logical_fallacies",
+          "mmlu:subject=machine_learning,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_machine_learning",
+          "mmlu:subject=management,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_management",
+          "mmlu:subject=marketing,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_marketing",
+          "mmlu:subject=medical_genetics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_medical_genetics",
+          "mmlu:subject=miscellaneous,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_miscellaneous",
+          "mmlu:subject=moral_disputes,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_disputes",
+          "mmlu:subject=moral_scenarios,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_moral_scenarios",
+          "mmlu:subject=nutrition,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_nutrition",
+          "mmlu:subject=philosophy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_philosophy",
+          "mmlu:subject=prehistory,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_prehistory",
+          "mmlu:subject=professional_accounting,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_accounting",
+          "mmlu:subject=professional_law,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_law",
+          "mmlu:subject=professional_medicine,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_medicine",
+          "mmlu:subject=professional_psychology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_professional_psychology",
+          "mmlu:subject=public_relations,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_public_relations",
+          "mmlu:subject=security_studies,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_security_studies",
+          "mmlu:subject=sociology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_sociology",
+          "mmlu:subject=us_foreign_policy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_us_foreign_policy",
+          "mmlu:subject=virology,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_virology",
+          "mmlu:subject=world_religions,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_world_religions"
+        ]
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_mmlu.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu.json"
+    }
+  ],
+  "name": "mmlu"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:abstract_algebra.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: abstract_algebra",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu&subgroup=subject%3A%20abstract_algebra&runSpecs=%5B%22mmlu%3Asubject%3Dabstract_algebra%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_abstract_algebra%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=abstract_algebra,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_abstract_algebra"
+        ]
+      },
+      {
+        "value": 0.23,
+        "description": "min=0.23, mean=0.23, max=0.23, sum=0.23 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06573188066482544,
+        "description": "min=0.066, mean=0.066, max=0.066, sum=0.066 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 358.76,
+        "description": "min=358.76, mean=358.76, max=358.76, sum=358.76 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_mmlu_subject:abstract_algebra.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:abstract_algebra.json"
+    }
+  ],
+  "name": "mmlu_subject:abstract_algebra"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:anatomy.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: anatomy",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu&subgroup=subject%3A%20anatomy&runSpecs=%5B%22mmlu%3Asubject%3Danatomy%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_anatomy%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=anatomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_anatomy"
+        ]
+      },
+      {
+        "value": 0.2962962962962963,
+        "description": "min=0.296, mean=0.296, max=0.296, sum=0.296 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.06830824922632288,
+        "description": "min=0.068, mean=0.068, max=0.068, sum=0.068 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 135.0,
+        "description": "min=135, mean=135, max=135, sum=135 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 352.02962962962965,
+        "description": "min=352.03, mean=352.03, max=352.03, sum=352.03 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_mmlu_subject:anatomy.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:anatomy.json"
+    }
+  ],
+  "name": "mmlu_subject:anatomy"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:astronomy.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: astronomy",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu&subgroup=subject%3A%20astronomy&runSpecs=%5B%22mmlu%3Asubject%3Dastronomy%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_astronomy%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=astronomy,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_astronomy"
+        ]
+      },
+      {
+        "value": 0.23684210526315788,
+        "description": "min=0.237, mean=0.237, max=0.237, sum=0.237 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.10126333958224247,
+        "description": "min=0.101, mean=0.101, max=0.101, sum=0.101 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 152.0,
+        "description": "min=152, mean=152, max=152, sum=152 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 598.4868421052631,
+        "description": "min=598.487, mean=598.487, max=598.487, sum=598.487 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_mmlu_subject:astronomy.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:astronomy.json"
+    }
+  ],
+  "name": "mmlu_subject:astronomy"
+}

mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:business_ethics.json

@@ -0,0 +1,145 @@
+{
+  "title": "subject: business_ethics",
+  "header": [
+    {
+      "value": "Model",
+      "markdown": false,
+      "metadata": {}
+    },
+    {
+      "value": "EM",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nExact match: Fraction of instances that the predicted output matches a correct reference exactly.",
+      "markdown": false,
+      "lower_is_better": false,
+      "metadata": {
+        "metric": "EM",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "Observed inference time (s)",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\nObserved inference runtime (s): Average observed time to process a request to the model (via an API, and thus depends on particular deployment).",
+      "markdown": false,
+      "lower_is_better": true,
+      "metadata": {
+        "metric": "Observed inference time (s)",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# eval",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# eval: Number of evaluation instances.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# eval",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# train",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# train: Number of training instances (e.g., in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "# train",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "truncated",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).",
+      "markdown": false,
+      "metadata": {
+        "metric": "truncated",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# prompt tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# prompt tokens: Number of tokens in the prompt.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# prompt tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    },
+    {
+      "value": "# output tokens",
+      "description": "The Massive Multitask Language Understanding (MMLU) benchmark for knowledge-intensive question answering across 57 domains [(Hendrycks et al., 2021)](https://arxiv.org/pdf/2009.03300.pdf).\n\n# output tokens: Actual number of output tokens.",
+      "markdown": false,
+      "metadata": {
+        "metric": "# output tokens",
+        "run_group": "MMLU All Subjects"
+      }
+    }
+  ],
+  "rows": [
+    [
+      {
+        "value": "EleutherAI/pythia-1.4b",
+        "description": "",
+        "href": "?group=mmlu&subgroup=subject%3A%20business_ethics&runSpecs=%5B%22mmlu%3Asubject%3Dbusiness_ethics%2Cmethod%3Dmultiple_choice_joint%2Cmodel%3DEleutherAI_pythia-1.4b%2Ceval_split%3Dtest%2Cgroups%3Dmmlu_business_ethics%22%5D",
+        "markdown": false,
+        "run_spec_names": [
+          "mmlu:subject=business_ethics,method=multiple_choice_joint,model=EleutherAI_pythia-1.4b,eval_split=test,groups=mmlu_business_ethics"
+        ]
+      },
+      {
+        "value": 0.29,
+        "description": "min=0.29, mean=0.29, max=0.29, sum=0.29 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 0.09688833236694336,
+        "description": "min=0.097, mean=0.097, max=0.097, sum=0.097 (1)",
+        "style": {
+          "font-weight": "bold"
+        },
+        "markdown": false
+      },
+      {
+        "value": 100.0,
+        "description": "min=100, mean=100, max=100, sum=100 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 5.0,
+        "description": "min=5, mean=5, max=5, sum=5 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 0.0,
+        "description": "min=0, mean=0, max=0, sum=0 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 585.05,
+        "description": "min=585.05, mean=585.05, max=585.05, sum=585.05 (1)",
+        "style": {},
+        "markdown": false
+      },
+      {
+        "value": 1.0,
+        "description": "min=1, mean=1, max=1, sum=1 (1)",
+        "style": {},
+        "markdown": false
+      }
+    ]
+  ],
+  "links": [
+    {
+      "text": "LaTeX",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/latex/mmlu_mmlu_subject:business_ethics.tex"
+    },
+    {
+      "text": "JSON",
+      "href": "benchmark_output/runs/mmlu_pythia-1.4b-step2000/groups/json/mmlu_mmlu_subject:business_ethics.json"
+    }
+  ],
+  "name": "mmlu_subject:business_ethics"
+}