Add files using upload-large-folder tool

evals/general_suite/base_full_v2/student/results_2026-07-17T11-10-28.190178.json

@@ -0,0 +1,333 @@
+{
+  "results": {
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 1319,
+      "exact_match,strict-match": 0.0,
+      "exact_match_stderr,strict-match": 0.0,
+      "exact_match,flexible-extract": 0.6944655041698257,
+      "exact_match_stderr,flexible-extract": 0.01268813407672688
+    },
+    "minerva_math500": {
+      "name": "minerva_math500",
+      "alias": "minerva_math500",
+      "sample_len": 500,
+      "exact_match,none": 0.1,
+      "exact_match_stderr,none": 0.013429844431075358,
+      "math_verify,none": 0.174,
+      "math_verify_stderr,none": 0.01697127125751611
+    },
+    "ifeval": {
+      "name": "ifeval",
+      "alias": "ifeval",
+      "sample_len": 541,
+      "prompt_level_strict_acc,none": 0.6155268022181146,
+      "prompt_level_strict_acc_stderr,none": 0.02093435763458476,
+      "inst_level_strict_acc,none": 0.7218225419664268,
+      "inst_level_strict_acc_stderr,none": "N/A",
+      "prompt_level_loose_acc,none": 0.6765249537892791,
+      "prompt_level_loose_acc_stderr,none": 0.020131003392118964,
+      "inst_level_loose_acc,none": 0.7757793764988009,
+      "inst_level_loose_acc_stderr,none": "N/A"
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    },
+    "ifeval": {
+      "task": "ifeval",
+      "dataset_path": "google/IFEval",
+      "test_split": "train",
+      "doc_to_text": "prompt",
+      "doc_to_target": 0,
+      "unsafe_code": false,
+      "process_results": "def process_results(doc, results):\n    inp = InputExample(\n        key=doc[\"key\"],\n        instruction_id_list=doc[\"instruction_id_list\"],\n        prompt=doc[\"prompt\"],\n        kwargs=doc[\"kwargs\"],\n    )\n    response = results[0]\n\n    out_strict = test_instruction_following_strict(inp, response)\n    out_loose = test_instruction_following_loose(inp, response)\n\n    return {\n        \"prompt_level_strict_acc\": out_strict.follow_all_instructions,\n        \"inst_level_strict_acc\": out_strict.follow_instruction_list,\n        \"prompt_level_loose_acc\": out_loose.follow_all_instructions,\n        \"inst_level_loose_acc\": out_loose.follow_instruction_list,\n    }\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "prompt",
+        "doc_to_choice": null,
+        "doc_to_target": 0,
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "prompt_level_strict_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_strict_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        },
+        {
+          "metric": "prompt_level_loose_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_loose_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 4.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/ifeval/ifeval.yaml"
+      }
+    },
+    "minerva_math500": {
+      "task": "minerva_math500",
+      "dataset_path": "HuggingFaceH4/MATH-500",
+      "dataset_name": "default",
+      "test_split": "test",
+      "process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n    def _process_doc(doc: dict) -> dict:\n        out_doc = {\n            \"problem\": doc[\"problem\"],\n            \"solution\": doc[\"solution\"],\n            \"answer\": normalize_final_answer(\n                remove_boxed(last_boxed_only_string(doc[\"solution\"]))\n            ),\n        }\n        if getattr(doc, \"few_shot\", None) is not None:\n            out_doc[\"few_shot\"] = True\n        return out_doc\n\n    return dataset.map(_process_doc)\n",
+      "doc_to_text": "def doc_to_text(doc: dict) -> str:\n    return \"Problem:\" + \"\\n\" + doc[\"problem\"] + \"\\n\\n\" + \"Solution:\"\n",
+      "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+      "unsafe_code": false,
+      "process_results": "def process_results(doc: dict, results: list[str]) -> dict[str, int]:\n    candidates = results[0]\n\n    unnormalized_answer = get_unnormalized_answer(candidates)\n    answer = normalize_final_answer(unnormalized_answer)\n\n    if is_equiv(answer, doc[\"answer\"]):\n        retval = 1\n    else:\n        retval = 0\n\n    # math_verify\n    _mvres = verify(\n        gold=parse(doc[\"solution\"]),\n        target=parse(candidates),\n    )\n    mathval = 1 if _mvres else 0\n\n    res = {\n        \"exact_match\": retval,\n        \"math_verify\": mathval,\n    }\n    return res\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": "<function process_docs at 0x7f02b6b2dbc0>",
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x7f02b6b2fa60>",
+        "doc_to_text": "<function doc_to_text at 0x7f02bc2f1940>",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 4,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "math_verify",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Problem:"
+        ],
+        "do_sample": false,
+        "temperature": 0.0
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/minerva_math/minerva_math500.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k_cot_zeroshot": 3.0,
+    "ifeval": 4.0,
+    "minerva_math500": 3.0
+  },
+  "n-shot": {
+    "gsm8k_cot_zeroshot": 0,
+    "ifeval": 0,
+    "minerva_math500": 4
+  },
+  "higher_is_better": {
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    },
+    "ifeval": {
+      "prompt_level_strict_acc": true,
+      "inst_level_strict_acc": true,
+      "prompt_level_loose_acc": true,
+      "inst_level_loose_acc": true
+    },
+    "minerva_math500": {
+      "exact_match": true,
+      "math_verify": true
+    }
+  },
+  "n-samples": {
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 1319
+    },
+    "minerva_math500": {
+      "original": 500,
+      "effective": 500
+    },
+    "ifeval": {
+      "original": 541,
+      "effective": 541
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784311549.3174064,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k_cot_zeroshot": "77ad80425a4cb85562b27f164a9c92977f775bb66a0ec4d374e4ee7b4656a7f5",
+    "minerva_math500": "5784cb060f783763ec8f8e234eef61589d348055c48e8adc0073445383258f85",
+    "ifeval": "bdcf10fc42fb3b39fd60af9d347a998b5bcb0a6782900c760b7dcc67ef0109d7"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "285.2433420190064"
+}

evals/general_suite/base_full_v2/student/results_2026-07-17T11-13-41.342240.json

@@ -0,0 +1,225 @@
+{
+  "results": {
+    "humaneval": {
+      "name": "humaneval",
+      "alias": "humaneval",
+      "sample_len": 164,
+      "pass@1,create_test": 0.3475609756097561,
+      "pass@1_stderr,create_test": 0.03729852575607224
+    },
+    "mbpp": {
+      "name": "mbpp",
+      "alias": "mbpp",
+      "sample_len": 500,
+      "pass_at_1,none": 0.302,
+      "pass_at_1_stderr,none": 0.020553269174209184
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "humaneval": {
+      "task": "humaneval",
+      "dataset_path": "openai/openai_humaneval",
+      "test_split": "test",
+      "doc_to_text": "{{prompt}}",
+      "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "{{prompt}}",
+        "doc_to_choice": null,
+        "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "def pass_at_k(references: list[str], predictions: list[list[str]], k: list[int] = None):\n    global compute_\n    assert k is not None\n    if isinstance(k, int):\n        k = [k]\n    res = compute_.compute(\n        references=references,\n        predictions=predictions,\n        k=k,\n    )\n    return res[0]\n",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "k": [
+            1
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "\nclass",
+          "\ndef",
+          "\n#",
+          "\nif",
+          "\nprint"
+        ],
+        "max_gen_toks": 1024,
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "create_test",
+          "filter": [
+            {
+              "function": "custom",
+              "filter_fn": "<function build_predictions at 0x76a40e2260c0>"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/completions",
+        "tokenizer": "allenai/OLMoE-1B-7B-0125-Instruct",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/humaneval/humaneval.yaml"
+      }
+    },
+    "mbpp": {
+      "task": "mbpp",
+      "dataset_path": "google-research-datasets/mbpp",
+      "dataset_name": "full",
+      "test_split": "test",
+      "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+      "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": "",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x76a4d6efb920>",
+        "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+        "doc_to_choice": null,
+        "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": ""
+      },
+      "num_fewshot": 3,
+      "metric_list": [
+        {
+          "metric": "def pass_at_1(\n    references: Union[str, list[str]], predictions: Union[str, list[list[str]]]\n) -> float:\n    if isinstance(references, str):\n        references = [references]\n    if isinstance(predictions[0], str):\n        predictions = [[p] for p in predictions]\n    return pass_at_k.compute(\n        references=references,\n        predictions=predictions,\n        k=[1],\n    )[0][\"pass@1\"]\n",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "[DONE]"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/completions",
+        "tokenizer": "allenai/OLMoE-1B-7B-0125-Instruct",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/mbpp/mbpp.yaml"
+      }
+    }
+  },
+  "versions": {
+    "humaneval": 1.0,
+    "mbpp": 1.0
+  },
+  "n-shot": {
+    "humaneval": 0,
+    "mbpp": 3
+  },
+  "higher_is_better": {
+    "humaneval": {
+      "pass_at_k": true,
+      "pass@1": true
+    },
+    "mbpp": {
+      "pass_at_1": true
+    }
+  },
+  "n-samples": {
+    "humaneval": {
+      "original": 164,
+      "effective": 164
+    },
+    "mbpp": {
+      "original": 500,
+      "effective": 500
+    }
+  },
+  "config": {
+    "model": "local-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8399/v1/completions",
+      "tokenizer": "allenai/OLMoE-1B-7B-0125-Instruct",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784311837.3935306,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "tokenizer_pad_token": [
+    "<pad>",
+    "50280"
+  ],
+  "tokenizer_eos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "tokenizer_bos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "eot_token_id": 50279,
+  "max_length": 2047,
+  "task_hashes": {
+    "humaneval": "c122632f0bdffdd0162cc8033a879dfe9d3b35e82c52eda6a767069e0a2beb40",
+    "mbpp": "62356f643ed511c895814c093ddb55451941156ba9febb31dc87c6e00d9723e3"
+  },
+  "model_source": "local-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": null,
+  "chat_template": null,
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "190.31868797099742"
+}

evals/general_suite/gsm8k_protocol/student/results_2026-07-17T10-25-18.320120.json

@@ -0,0 +1,412 @@
+{
+  "results": {
+    "gsm8k": {
+      "name": "gsm8k",
+      "alias": "gsm8k",
+      "sample_len": 100,
+      "exact_match,strict-match": 0.71,
+      "exact_match_stderr,strict-match": 0.04560480215720684,
+      "exact_match,flexible-extract": 0.74,
+      "exact_match_stderr,flexible-extract": 0.04408440022768077
+    },
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 100,
+      "exact_match,strict-match": 0.0,
+      "exact_match_stderr,strict-match": 0.0,
+      "exact_match,flexible-extract": 0.7,
+      "exact_match_stderr,flexible-extract": 0.046056618647183814
+    },
+    "gsm8k_cot": {
+      "name": "gsm8k_cot",
+      "alias": "gsm8k_cot",
+      "sample_len": 100,
+      "exact_match,strict-match": 0.69,
+      "exact_match_stderr,strict-match": 0.04648231987117316,
+      "exact_match,flexible-extract": 0.7,
+      "exact_match_stderr,flexible-extract": 0.046056618647183814
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k": {
+      "task": "gsm8k",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Question: {{question}}\nAnswer:",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Question: {{question}}\nAnswer:",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 5,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Question:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false,
+        "temperature": 0.0
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "#### (\\-?[0-9\\.\\,]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+        "num_concurrent": 32,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k.yaml"
+      }
+    },
+    "gsm8k_cot": {
+      "task": "gsm8k_cot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "test_split": "test",
+      "doc_to_text": "Q: {{question}}\nA:",
+      "doc_to_target": "{{answer.split('####')[-1].strip() if answer is defined else target}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": [
+          {
+            "question": "There are 15 trees in the grove. Grove workers will plant trees in the grove today. After they are done, there will be 21 trees. How many trees did the grove workers plant today?",
+            "target": "There are 15 trees originally. Then there were 21 trees after some more were planted. So there must have been 21 - 15 = 6. The answer is 6."
+          },
+          {
+            "question": "If there are 3 cars in the parking lot and 2 more cars arrive, how many cars are in the parking lot?",
+            "target": "There are originally 3 cars. 2 more cars arrive. 3 + 2 = 5. The answer is 5."
+          },
+          {
+            "question": "Leah had 32 chocolates and her sister had 42. If they ate 35, how many pieces do they have left in total?",
+            "target": "Originally, Leah had 32 chocolates. Her sister had 42. So in total they had 32 + 42 = 74. After eating 35, they had 74 - 35 = 39. The answer is 39."
+          },
+          {
+            "question": "Jason had 20 lollipops. He gave Denny some lollipops. Now Jason has 12 lollipops. How many lollipops did Jason give to Denny?",
+            "target": "Jason started with 20 lollipops. Then he had 12 after giving some to Denny. So he gave Denny 20 - 12 = 8. The answer is 8."
+          },
+          {
+            "question": "Shawn has five toys. For Christmas, he got two toys each from his mom and dad. How many toys does he have now?",
+            "target": "Shawn started with 5 toys. If he got 2 toys each from his mom and dad, then that is 4 more toys. 5 + 4 = 9. The answer is 9."
+          },
+          {
+            "question": "There were nine computers in the server room. Five more computers were installed each day, from monday to thursday. How many computers are now in the server room?",
+            "target": "There were originally 9 computers. For each of 4 days, 5 more computers were added. So 5 * 4 = 20 computers were added. 9 + 20 is 29. The answer is 29."
+          },
+          {
+            "question": "Michael had 58 golf balls. On tuesday, he lost 23 golf balls. On wednesday, he lost 2 more. How many golf balls did he have at the end of wednesday?",
+            "target": "Michael started with 58 golf balls. After losing 23 on tuesday, he had 58 - 23 = 35. After losing 2 more, he had 35 - 2 = 33 golf balls. The answer is 33."
+          },
+          {
+            "question": "Olivia has $23. She bought five bagels for $3 each. How much money does she have left?",
+            "target": "Olivia had 23 dollars. 5 bagels for 3 dollars each will be 5 x 3 = 15 dollars. So she has 23 - 15 dollars left. 23 - 15 is 8. The answer is 8."
+          }
+        ],
+        "doc_to_text": "Q: {{question}}\nA:",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer.split('####')[-1].strip() if answer is defined else target}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 8,
+      "metric_list": [
+        {
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "metric": "exact_match",
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "do_sample": false,
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ]
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ],
+          "name": "strict-match"
+        },
+        {
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ],
+          "name": "flexible-extract"
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+        "num_concurrent": 32,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot.yaml"
+      }
+    },
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+        "num_concurrent": 32,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k": 3.0,
+    "gsm8k_cot": 3.0,
+    "gsm8k_cot_zeroshot": 3.0
+  },
+  "n-shot": {
+    "gsm8k": 5,
+    "gsm8k_cot": 8,
+    "gsm8k_cot_zeroshot": 0
+  },
+  "higher_is_better": {
+    "gsm8k": {
+      "exact_match": true
+    },
+    "gsm8k_cot": {
+      "exact_match": true
+    },
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    }
+  },
+  "n-samples": {
+    "gsm8k": {
+      "original": 1319,
+      "effective": 100
+    },
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 100
+    },
+    "gsm8k_cot": {
+      "original": 1319,
+      "effective": 100
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8399/v1/chat/completions",
+      "num_concurrent": 32,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": 100.0,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784309081.761174,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k": "8d403e0ebd80e62a0057ac95d5dacc4e0426634748669111fd499bbe4cc2e9b1",
+    "gsm8k_cot_zeroshot": "68f0c791c6679e4867dc4535fe7469470d56e6de1daa8d170b515a68fd7de893",
+    "gsm8k_cot": "c037f4349978481ffff602dca0c0817aad9dc51b24ad156c21464ed9ffa327d3"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "42.89717419000226"
+}

evals/general_suite/healed/reap_keep25_s1224/student/results_2026-07-17T17-42-23.472237.json

@@ -0,0 +1,337 @@
+{
+  "results": {
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 1319,
+      "exact_match,strict-match": 0.00530705079605762,
+      "exact_match_stderr,strict-match": 0.0020013057209480522,
+      "exact_match,flexible-extract": 0.08567096285064443,
+      "exact_match_stderr,flexible-extract": 0.007709218855882765
+    },
+    "minerva_math500": {
+      "name": "minerva_math500",
+      "alias": "minerva_math500",
+      "sample_len": 500,
+      "exact_match,none": 0.026,
+      "exact_match_stderr,none": 0.007123868260853507,
+      "math_verify,none": 0.058,
+      "math_verify_stderr,none": 0.010463793860924815
+    },
+    "ifeval": {
+      "name": "ifeval",
+      "alias": "ifeval",
+      "sample_len": 541,
+      "prompt_level_strict_acc,none": 0.3142329020332717,
+      "prompt_level_strict_acc_stderr,none": 0.019976402065150656,
+      "inst_level_strict_acc,none": 0.45443645083932854,
+      "inst_level_strict_acc_stderr,none": "N/A",
+      "prompt_level_loose_acc,none": 0.3419593345656192,
+      "prompt_level_loose_acc_stderr,none": 0.020413464513923615,
+      "inst_level_loose_acc,none": 0.47961630695443647,
+      "inst_level_loose_acc_stderr,none": "N/A"
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8420/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    },
+    "ifeval": {
+      "task": "ifeval",
+      "dataset_path": "google/IFEval",
+      "test_split": "train",
+      "doc_to_text": "prompt",
+      "doc_to_target": 0,
+      "unsafe_code": false,
+      "process_results": "def process_results(doc, results):\n    inp = InputExample(\n        key=doc[\"key\"],\n        instruction_id_list=doc[\"instruction_id_list\"],\n        prompt=doc[\"prompt\"],\n        kwargs=doc[\"kwargs\"],\n    )\n    response = results[0]\n\n    out_strict = test_instruction_following_strict(inp, response)\n    out_loose = test_instruction_following_loose(inp, response)\n\n    return {\n        \"prompt_level_strict_acc\": out_strict.follow_all_instructions,\n        \"inst_level_strict_acc\": out_strict.follow_instruction_list,\n        \"prompt_level_loose_acc\": out_loose.follow_all_instructions,\n        \"inst_level_loose_acc\": out_loose.follow_instruction_list,\n    }\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "prompt",
+        "doc_to_choice": null,
+        "doc_to_target": 0,
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "prompt_level_strict_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_strict_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        },
+        {
+          "metric": "prompt_level_loose_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_loose_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 4.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8420/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/ifeval/ifeval.yaml"
+      }
+    },
+    "minerva_math500": {
+      "task": "minerva_math500",
+      "dataset_path": "HuggingFaceH4/MATH-500",
+      "dataset_name": "default",
+      "test_split": "test",
+      "process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n    def _process_doc(doc: dict) -> dict:\n        out_doc = {\n            \"problem\": doc[\"problem\"],\n            \"solution\": doc[\"solution\"],\n            \"answer\": normalize_final_answer(\n                remove_boxed(last_boxed_only_string(doc[\"solution\"]))\n            ),\n        }\n        if getattr(doc, \"few_shot\", None) is not None:\n            out_doc[\"few_shot\"] = True\n        return out_doc\n\n    return dataset.map(_process_doc)\n",
+      "doc_to_text": "def doc_to_text(doc: dict) -> str:\n    return \"Problem:\" + \"\\n\" + doc[\"problem\"] + \"\\n\\n\" + \"Solution:\"\n",
+      "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+      "unsafe_code": false,
+      "process_results": "def process_results(doc: dict, results: list[str]) -> dict[str, int]:\n    candidates = results[0]\n\n    unnormalized_answer = get_unnormalized_answer(candidates)\n    answer = normalize_final_answer(unnormalized_answer)\n\n    if is_equiv(answer, doc[\"answer\"]):\n        retval = 1\n    else:\n        retval = 0\n\n    # math_verify\n    _mvres = verify(\n        gold=parse(doc[\"solution\"]),\n        target=parse(candidates),\n    )\n    mathval = 1 if _mvres else 0\n\n    res = {\n        \"exact_match\": retval,\n        \"math_verify\": mathval,\n    }\n    return res\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": "<function process_docs at 0x7d9c1442db20>",
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x7d9c1442f9c0>",
+        "doc_to_text": "<function doc_to_text at 0x7d9c15bf9800>",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 4,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "math_verify",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Problem:"
+        ],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8420/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/minerva_math/minerva_math500.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k_cot_zeroshot": 3.0,
+    "ifeval": 4.0,
+    "minerva_math500": 3.0
+  },
+  "n-shot": {
+    "gsm8k_cot_zeroshot": 0,
+    "ifeval": 0,
+    "minerva_math500": 4
+  },
+  "higher_is_better": {
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    },
+    "ifeval": {
+      "prompt_level_strict_acc": true,
+      "inst_level_strict_acc": true,
+      "prompt_level_loose_acc": true,
+      "inst_level_loose_acc": true
+    },
+    "minerva_math500": {
+      "exact_match": true,
+      "math_verify": true
+    }
+  },
+  "n-samples": {
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 1319
+    },
+    "minerva_math500": {
+      "original": 500,
+      "effective": 500
+    },
+    "ifeval": {
+      "original": 541,
+      "effective": 541
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8420/v1/chat/completions",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {
+      "max_gen_toks": 1280
+    },
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784334723.3197055,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k_cot_zeroshot": "77ad80425a4cb85562b27f164a9c92977f775bb66a0ec4d374e4ee7b4656a7f5",
+    "minerva_math500": "5784cb060f783763ec8f8e234eef61589d348055c48e8adc0073445383258f85",
+    "ifeval": "bdcf10fc42fb3b39fd60af9d347a998b5bcb0a6782900c760b7dcc67ef0109d7"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "626.5327842670085"
+}

evals/general_suite/healed/reap_keep25_s1224/student/results_2026-07-17T17-45-52.249309.json

@@ -0,0 +1,225 @@
+{
+  "results": {
+    "humaneval": {
+      "name": "humaneval",
+      "alias": "humaneval",
+      "sample_len": 164,
+      "pass@1,create_test": 0.042682926829268296,
+      "pass@1_stderr,create_test": 0.015832926510304685
+    },
+    "mbpp": {
+      "name": "mbpp",
+      "alias": "mbpp",
+      "sample_len": 500,
+      "pass_at_1,none": 0.106,
+      "pass_at_1_stderr,none": 0.01378070446788774
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "humaneval": {
+      "task": "humaneval",
+      "dataset_path": "openai/openai_humaneval",
+      "test_split": "test",
+      "doc_to_text": "{{prompt}}",
+      "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "{{prompt}}",
+        "doc_to_choice": null,
+        "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "def pass_at_k(references: list[str], predictions: list[list[str]], k: list[int] = None):\n    global compute_\n    assert k is not None\n    if isinstance(k, int):\n        k = [k]\n    res = compute_.compute(\n        references=references,\n        predictions=predictions,\n        k=k,\n    )\n    return res[0]\n",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "k": [
+            1
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "\nclass",
+          "\ndef",
+          "\n#",
+          "\nif",
+          "\nprint"
+        ],
+        "max_gen_toks": 1024,
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "create_test",
+          "filter": [
+            {
+              "function": "custom",
+              "filter_fn": "<function build_predictions at 0x7c9507f6eca0>"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8420/v1/completions",
+        "tokenizer": "outputs/healed/grid_general/reap_keep25_s1224/step0150",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/humaneval/humaneval.yaml"
+      }
+    },
+    "mbpp": {
+      "task": "mbpp",
+      "dataset_path": "google-research-datasets/mbpp",
+      "dataset_name": "full",
+      "test_split": "test",
+      "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+      "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": "",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x7c95d10ff9c0>",
+        "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+        "doc_to_choice": null,
+        "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": ""
+      },
+      "num_fewshot": 3,
+      "metric_list": [
+        {
+          "metric": "def pass_at_1(\n    references: Union[str, list[str]], predictions: Union[str, list[list[str]]]\n) -> float:\n    if isinstance(references, str):\n        references = [references]\n    if isinstance(predictions[0], str):\n        predictions = [[p] for p in predictions]\n    return pass_at_k.compute(\n        references=references,\n        predictions=predictions,\n        k=[1],\n    )[0][\"pass@1\"]\n",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "[DONE]"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8420/v1/completions",
+        "tokenizer": "outputs/healed/grid_general/reap_keep25_s1224/step0150",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/mbpp/mbpp.yaml"
+      }
+    }
+  },
+  "versions": {
+    "humaneval": 1.0,
+    "mbpp": 1.0
+  },
+  "n-shot": {
+    "humaneval": 0,
+    "mbpp": 3
+  },
+  "higher_is_better": {
+    "humaneval": {
+      "pass_at_k": true,
+      "pass@1": true
+    },
+    "mbpp": {
+      "pass_at_1": true
+    }
+  },
+  "n-samples": {
+    "humaneval": {
+      "original": 164,
+      "effective": 164
+    },
+    "mbpp": {
+      "original": 500,
+      "effective": 500
+    }
+  },
+  "config": {
+    "model": "local-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8420/v1/completions",
+      "tokenizer": "outputs/healed/grid_general/reap_keep25_s1224/step0150",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784335352.5630722,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "tokenizer_pad_token": [
+    "<pad>",
+    "50280"
+  ],
+  "tokenizer_eos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "tokenizer_bos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "eot_token_id": 50279,
+  "max_length": 2047,
+  "task_hashes": {
+    "humaneval": "c122632f0bdffdd0162cc8033a879dfe9d3b35e82c52eda6a767069e0a2beb40",
+    "mbpp": "62356f643ed511c895814c093ddb55451941156ba9febb31dc87c6e00d9723e3"
+  },
+  "model_source": "local-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": null,
+  "chat_template": null,
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "206.03511413600063"
+}

evals/general_suite/healed/reap_keep25_s1225/student/results_2026-07-17T17-25-02.868375.json

@@ -0,0 +1,337 @@
+{
+  "results": {
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 1319,
+      "exact_match,strict-match": 0.004548900682335102,
+      "exact_match_stderr,strict-match": 0.0018535550440036193,
+      "exact_match,flexible-extract": 0.09173616376042457,
+      "exact_match_stderr,flexible-extract": 0.007950942148339347
+    },
+    "minerva_math500": {
+      "name": "minerva_math500",
+      "alias": "minerva_math500",
+      "sample_len": 500,
+      "exact_match,none": 0.016,
+      "exact_match_stderr,none": 0.005617036781471955,
+      "math_verify,none": 0.034,
+      "math_verify_stderr,none": 0.008112930375552172
+    },
+    "ifeval": {
+      "name": "ifeval",
+      "alias": "ifeval",
+      "sample_len": 541,
+      "prompt_level_strict_acc,none": 0.3142329020332717,
+      "prompt_level_strict_acc_stderr,none": 0.019976402065150656,
+      "inst_level_strict_acc,none": 0.44844124700239807,
+      "inst_level_strict_acc_stderr,none": "N/A",
+      "prompt_level_loose_acc,none": 0.3419593345656192,
+      "prompt_level_loose_acc_stderr,none": 0.020413464513923615,
+      "inst_level_loose_acc,none": 0.4784172661870504,
+      "inst_level_loose_acc_stderr,none": "N/A"
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    },
+    "ifeval": {
+      "task": "ifeval",
+      "dataset_path": "google/IFEval",
+      "test_split": "train",
+      "doc_to_text": "prompt",
+      "doc_to_target": 0,
+      "unsafe_code": false,
+      "process_results": "def process_results(doc, results):\n    inp = InputExample(\n        key=doc[\"key\"],\n        instruction_id_list=doc[\"instruction_id_list\"],\n        prompt=doc[\"prompt\"],\n        kwargs=doc[\"kwargs\"],\n    )\n    response = results[0]\n\n    out_strict = test_instruction_following_strict(inp, response)\n    out_loose = test_instruction_following_loose(inp, response)\n\n    return {\n        \"prompt_level_strict_acc\": out_strict.follow_all_instructions,\n        \"inst_level_strict_acc\": out_strict.follow_instruction_list,\n        \"prompt_level_loose_acc\": out_loose.follow_all_instructions,\n        \"inst_level_loose_acc\": out_loose.follow_instruction_list,\n    }\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "prompt",
+        "doc_to_choice": null,
+        "doc_to_target": 0,
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "prompt_level_strict_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_strict_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        },
+        {
+          "metric": "prompt_level_loose_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_loose_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 4.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/ifeval/ifeval.yaml"
+      }
+    },
+    "minerva_math500": {
+      "task": "minerva_math500",
+      "dataset_path": "HuggingFaceH4/MATH-500",
+      "dataset_name": "default",
+      "test_split": "test",
+      "process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n    def _process_doc(doc: dict) -> dict:\n        out_doc = {\n            \"problem\": doc[\"problem\"],\n            \"solution\": doc[\"solution\"],\n            \"answer\": normalize_final_answer(\n                remove_boxed(last_boxed_only_string(doc[\"solution\"]))\n            ),\n        }\n        if getattr(doc, \"few_shot\", None) is not None:\n            out_doc[\"few_shot\"] = True\n        return out_doc\n\n    return dataset.map(_process_doc)\n",
+      "doc_to_text": "def doc_to_text(doc: dict) -> str:\n    return \"Problem:\" + \"\\n\" + doc[\"problem\"] + \"\\n\\n\" + \"Solution:\"\n",
+      "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+      "unsafe_code": false,
+      "process_results": "def process_results(doc: dict, results: list[str]) -> dict[str, int]:\n    candidates = results[0]\n\n    unnormalized_answer = get_unnormalized_answer(candidates)\n    answer = normalize_final_answer(unnormalized_answer)\n\n    if is_equiv(answer, doc[\"answer\"]):\n        retval = 1\n    else:\n        retval = 0\n\n    # math_verify\n    _mvres = verify(\n        gold=parse(doc[\"solution\"]),\n        target=parse(candidates),\n    )\n    mathval = 1 if _mvres else 0\n\n    res = {\n        \"exact_match\": retval,\n        \"math_verify\": mathval,\n    }\n    return res\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": "<function process_docs at 0x73cea223db20>",
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x73cea223f9c0>",
+        "doc_to_text": "<function doc_to_text at 0x73cea3809800>",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 4,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "math_verify",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Problem:"
+        ],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/minerva_math/minerva_math500.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k_cot_zeroshot": 3.0,
+    "ifeval": 4.0,
+    "minerva_math500": 3.0
+  },
+  "n-shot": {
+    "gsm8k_cot_zeroshot": 0,
+    "ifeval": 0,
+    "minerva_math500": 4
+  },
+  "higher_is_better": {
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    },
+    "ifeval": {
+      "prompt_level_strict_acc": true,
+      "inst_level_strict_acc": true,
+      "prompt_level_loose_acc": true,
+      "inst_level_loose_acc": true
+    },
+    "minerva_math500": {
+      "exact_match": true,
+      "math_verify": true
+    }
+  },
+  "n-samples": {
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 1319
+    },
+    "minerva_math500": {
+      "original": 500,
+      "effective": 500
+    },
+    "ifeval": {
+      "original": 541,
+      "effective": 541
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {
+      "max_gen_toks": 1280
+    },
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784333857.3417559,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k_cot_zeroshot": "77ad80425a4cb85562b27f164a9c92977f775bb66a0ec4d374e4ee7b4656a7f5",
+    "minerva_math500": "5784cb060f783763ec8f8e234eef61589d348055c48e8adc0073445383258f85",
+    "ifeval": "bdcf10fc42fb3b39fd60af9d347a998b5bcb0a6782900c760b7dcc67ef0109d7"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "451.9168644050078"
+}

evals/general_suite/healed/reap_keep25_s1225/student/results_2026-07-17T17-29-11.051064.json

@@ -0,0 +1,225 @@
+{
+  "results": {
+    "humaneval": {
+      "name": "humaneval",
+      "alias": "humaneval",
+      "sample_len": 164,
+      "pass@1,create_test": 0.04878048780487805,
+      "pass@1_stderr,create_test": 0.016872119488648706
+    },
+    "mbpp": {
+      "name": "mbpp",
+      "alias": "mbpp",
+      "sample_len": 500,
+      "pass_at_1,none": 0.112,
+      "pass_at_1_stderr,none": 0.014117741364768066
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "humaneval": {
+      "task": "humaneval",
+      "dataset_path": "openai/openai_humaneval",
+      "test_split": "test",
+      "doc_to_text": "{{prompt}}",
+      "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "{{prompt}}",
+        "doc_to_choice": null,
+        "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "def pass_at_k(references: list[str], predictions: list[list[str]], k: list[int] = None):\n    global compute_\n    assert k is not None\n    if isinstance(k, int):\n        k = [k]\n    res = compute_.compute(\n        references=references,\n        predictions=predictions,\n        k=k,\n    )\n    return res[0]\n",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "k": [
+            1
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "\nclass",
+          "\ndef",
+          "\n#",
+          "\nif",
+          "\nprint"
+        ],
+        "max_gen_toks": 1024,
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "create_test",
+          "filter": [
+            {
+              "function": "custom",
+              "filter_fn": "<function build_predictions at 0x775f05d76f20>"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/completions",
+        "tokenizer": "outputs/healed/grid_general/reap_keep25_s1225/step0150",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/humaneval/humaneval.yaml"
+      }
+    },
+    "mbpp": {
+      "task": "mbpp",
+      "dataset_path": "google-research-datasets/mbpp",
+      "dataset_name": "full",
+      "test_split": "test",
+      "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+      "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": "",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x775fcf12f9c0>",
+        "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+        "doc_to_choice": null,
+        "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": ""
+      },
+      "num_fewshot": 3,
+      "metric_list": [
+        {
+          "metric": "def pass_at_1(\n    references: Union[str, list[str]], predictions: Union[str, list[list[str]]]\n) -> float:\n    if isinstance(references, str):\n        references = [references]\n    if isinstance(predictions[0], str):\n        predictions = [[p] for p in predictions]\n    return pass_at_k.compute(\n        references=references,\n        predictions=predictions,\n        k=[1],\n    )[0][\"pass@1\"]\n",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "[DONE]"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/completions",
+        "tokenizer": "outputs/healed/grid_general/reap_keep25_s1225/step0150",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/mbpp/mbpp.yaml"
+      }
+    }
+  },
+  "versions": {
+    "humaneval": 1.0,
+    "mbpp": 1.0
+  },
+  "n-shot": {
+    "humaneval": 0,
+    "mbpp": 3
+  },
+  "higher_is_better": {
+    "humaneval": {
+      "pass_at_k": true,
+      "pass@1": true
+    },
+    "mbpp": {
+      "pass_at_1": true
+    }
+  },
+  "n-samples": {
+    "humaneval": {
+      "original": 164,
+      "effective": 164
+    },
+    "mbpp": {
+      "original": 500,
+      "effective": 500
+    }
+  },
+  "config": {
+    "model": "local-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/completions",
+      "tokenizer": "outputs/healed/grid_general/reap_keep25_s1225/step0150",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784334311.9486935,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "tokenizer_pad_token": [
+    "<pad>",
+    "50280"
+  ],
+  "tokenizer_eos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "tokenizer_bos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "eot_token_id": 50279,
+  "max_length": 2047,
+  "task_hashes": {
+    "humaneval": "c122632f0bdffdd0162cc8033a879dfe9d3b35e82c52eda6a767069e0a2beb40",
+    "mbpp": "62356f643ed511c895814c093ddb55451941156ba9febb31dc87c6e00d9723e3"
+  },
+  "model_source": "local-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": null,
+  "chat_template": null,
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "245.4696692580037"
+}

evals/general_suite/healed/reap_keep50_s1224_long500_step150/student/results_2026-07-19T22-55-09.344078.json

@@ -0,0 +1,337 @@
+{
+  "results": {
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 1319,
+      "exact_match,strict-match": 0.003032600454890068,
+      "exact_match_stderr,strict-match": 0.001514573561224549,
+      "exact_match,flexible-extract": 0.3237300985595148,
+      "exact_match_stderr,flexible-extract": 0.012888247397371138
+    },
+    "minerva_math500": {
+      "name": "minerva_math500",
+      "alias": "minerva_math500",
+      "sample_len": 500,
+      "exact_match,none": 0.09,
+      "exact_match_stderr,none": 0.012811255071733842,
+      "math_verify,none": 0.126,
+      "math_verify_stderr,none": 0.01485561775078758
+    },
+    "ifeval": {
+      "name": "ifeval",
+      "alias": "ifeval",
+      "sample_len": 541,
+      "prompt_level_strict_acc,none": 0.5009242144177449,
+      "prompt_level_strict_acc_stderr,none": 0.02151653738784254,
+      "inst_level_strict_acc,none": 0.6187050359712231,
+      "inst_level_strict_acc_stderr,none": "N/A",
+      "prompt_level_loose_acc,none": 0.5286506469500925,
+      "prompt_level_loose_acc_stderr,none": 0.021481220930080487,
+      "inst_level_loose_acc,none": 0.6474820143884892,
+      "inst_level_loose_acc_stderr,none": "N/A"
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    },
+    "ifeval": {
+      "task": "ifeval",
+      "dataset_path": "google/IFEval",
+      "test_split": "train",
+      "doc_to_text": "prompt",
+      "doc_to_target": 0,
+      "unsafe_code": false,
+      "process_results": "def process_results(doc, results):\n    inp = InputExample(\n        key=doc[\"key\"],\n        instruction_id_list=doc[\"instruction_id_list\"],\n        prompt=doc[\"prompt\"],\n        kwargs=doc[\"kwargs\"],\n    )\n    response = results[0]\n\n    out_strict = test_instruction_following_strict(inp, response)\n    out_loose = test_instruction_following_loose(inp, response)\n\n    return {\n        \"prompt_level_strict_acc\": out_strict.follow_all_instructions,\n        \"inst_level_strict_acc\": out_strict.follow_instruction_list,\n        \"prompt_level_loose_acc\": out_loose.follow_all_instructions,\n        \"inst_level_loose_acc\": out_loose.follow_instruction_list,\n    }\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "prompt",
+        "doc_to_choice": null,
+        "doc_to_target": 0,
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "prompt_level_strict_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_strict_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        },
+        {
+          "metric": "prompt_level_loose_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_loose_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 4.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/ifeval/ifeval.yaml"
+      }
+    },
+    "minerva_math500": {
+      "task": "minerva_math500",
+      "dataset_path": "HuggingFaceH4/MATH-500",
+      "dataset_name": "default",
+      "test_split": "test",
+      "process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n    def _process_doc(doc: dict) -> dict:\n        out_doc = {\n            \"problem\": doc[\"problem\"],\n            \"solution\": doc[\"solution\"],\n            \"answer\": normalize_final_answer(\n                remove_boxed(last_boxed_only_string(doc[\"solution\"]))\n            ),\n        }\n        if getattr(doc, \"few_shot\", None) is not None:\n            out_doc[\"few_shot\"] = True\n        return out_doc\n\n    return dataset.map(_process_doc)\n",
+      "doc_to_text": "def doc_to_text(doc: dict) -> str:\n    return \"Problem:\" + \"\\n\" + doc[\"problem\"] + \"\\n\\n\" + \"Solution:\"\n",
+      "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+      "unsafe_code": false,
+      "process_results": "def process_results(doc: dict, results: list[str]) -> dict[str, int]:\n    candidates = results[0]\n\n    unnormalized_answer = get_unnormalized_answer(candidates)\n    answer = normalize_final_answer(unnormalized_answer)\n\n    if is_equiv(answer, doc[\"answer\"]):\n        retval = 1\n    else:\n        retval = 0\n\n    # math_verify\n    _mvres = verify(\n        gold=parse(doc[\"solution\"]),\n        target=parse(candidates),\n    )\n    mathval = 1 if _mvres else 0\n\n    res = {\n        \"exact_match\": retval,\n        \"math_verify\": mathval,\n    }\n    return res\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": "<function process_docs at 0x71680b81db20>",
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x71680b81f9c0>",
+        "doc_to_text": "<function doc_to_text at 0x716810fe58a0>",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 4,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "math_verify",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Problem:"
+        ],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/minerva_math/minerva_math500.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k_cot_zeroshot": 3.0,
+    "ifeval": 4.0,
+    "minerva_math500": 3.0
+  },
+  "n-shot": {
+    "gsm8k_cot_zeroshot": 0,
+    "ifeval": 0,
+    "minerva_math500": 4
+  },
+  "higher_is_better": {
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    },
+    "ifeval": {
+      "prompt_level_strict_acc": true,
+      "inst_level_strict_acc": true,
+      "prompt_level_loose_acc": true,
+      "inst_level_loose_acc": true
+    },
+    "minerva_math500": {
+      "exact_match": true,
+      "math_verify": true
+    }
+  },
+  "n-samples": {
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 1319
+    },
+    "minerva_math500": {
+      "original": 500,
+      "effective": 500
+    },
+    "ifeval": {
+      "original": 541,
+      "effective": 541
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {
+      "max_gen_toks": 1280
+    },
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "4cd474a",
+  "date": 1784526569.364177,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.11\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k_cot_zeroshot": "77ad80425a4cb85562b27f164a9c92977f775bb66a0ec4d374e4ee7b4656a7f5",
+    "minerva_math500": "5784cb060f783763ec8f8e234eef61589d348055c48e8adc0073445383258f85",
+    "ifeval": "bdcf10fc42fb3b39fd60af9d347a998b5bcb0a6782900c760b7dcc67ef0109d7"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "346.35699901598855"
+}

evals/general_suite/healed/reap_keep50_s1224_long500_step150/student/results_2026-07-19T22-58-18.278433.json

@@ -0,0 +1,225 @@
+{
+  "results": {
+    "humaneval": {
+      "name": "humaneval",
+      "alias": "humaneval",
+      "sample_len": 164,
+      "pass@1,create_test": 0.21951219512195122,
+      "pass@1_stderr,create_test": 0.032420416133953876
+    },
+    "mbpp": {
+      "name": "mbpp",
+      "alias": "mbpp",
+      "sample_len": 500,
+      "pass_at_1,none": 0.228,
+      "pass_at_1_stderr,none": 0.0187813065293632
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "humaneval": {
+      "task": "humaneval",
+      "dataset_path": "openai/openai_humaneval",
+      "test_split": "test",
+      "doc_to_text": "{{prompt}}",
+      "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "{{prompt}}",
+        "doc_to_choice": null,
+        "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "def pass_at_k(references: list[str], predictions: list[list[str]], k: list[int] = None):\n    global compute_\n    assert k is not None\n    if isinstance(k, int):\n        k = [k]\n    res = compute_.compute(\n        references=references,\n        predictions=predictions,\n        k=k,\n    )\n    return res[0]\n",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "k": [
+            1
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "\nclass",
+          "\ndef",
+          "\n#",
+          "\nif",
+          "\nprint"
+        ],
+        "max_gen_toks": 1024,
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "create_test",
+          "filter": [
+            {
+              "function": "custom",
+              "filter_fn": "<function build_predictions at 0x73240ec5eca0>"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/completions",
+        "tokenizer": "outputs/healed/grid_general_fairness/reap_keep50_s1224_long500/step0150",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/humaneval/humaneval.yaml"
+      }
+    },
+    "mbpp": {
+      "task": "mbpp",
+      "dataset_path": "google-research-datasets/mbpp",
+      "dataset_name": "full",
+      "test_split": "test",
+      "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+      "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": "",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x7324d7afb9c0>",
+        "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+        "doc_to_choice": null,
+        "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": ""
+      },
+      "num_fewshot": 3,
+      "metric_list": [
+        {
+          "metric": "def pass_at_1(\n    references: Union[str, list[str]], predictions: Union[str, list[list[str]]]\n) -> float:\n    if isinstance(references, str):\n        references = [references]\n    if isinstance(predictions[0], str):\n        predictions = [[p] for p in predictions]\n    return pass_at_k.compute(\n        references=references,\n        predictions=predictions,\n        k=[1],\n    )[0][\"pass@1\"]\n",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "[DONE]"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/completions",
+        "tokenizer": "outputs/healed/grid_general_fairness/reap_keep50_s1224_long500/step0150",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/mbpp/mbpp.yaml"
+      }
+    }
+  },
+  "versions": {
+    "humaneval": 1.0,
+    "mbpp": 1.0
+  },
+  "n-shot": {
+    "humaneval": 0,
+    "mbpp": 3
+  },
+  "higher_is_better": {
+    "humaneval": {
+      "pass_at_k": true,
+      "pass@1": true
+    },
+    "mbpp": {
+      "pass_at_1": true
+    }
+  },
+  "n-samples": {
+    "humaneval": {
+      "original": 164,
+      "effective": 164
+    },
+    "mbpp": {
+      "original": 500,
+      "effective": 500
+    }
+  },
+  "config": {
+    "model": "local-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/completions",
+      "tokenizer": "outputs/healed/grid_general_fairness/reap_keep50_s1224_long500/step0150",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "4cd474a",
+  "date": 1784526918.8328285,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.11\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "tokenizer_pad_token": [
+    "<pad>",
+    "50280"
+  ],
+  "tokenizer_eos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "tokenizer_bos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "eot_token_id": 50279,
+  "max_length": 2047,
+  "task_hashes": {
+    "humaneval": "c122632f0bdffdd0162cc8033a879dfe9d3b35e82c52eda6a767069e0a2beb40",
+    "mbpp": "62356f643ed511c895814c093ddb55451941156ba9febb31dc87c6e00d9723e3"
+  },
+  "model_source": "local-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": null,
+  "chat_template": null,
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "185.7945135810005"
+}

evals/general_suite/healed/reap_keep50_s1224_long500_step500/student/results_2026-07-19T23-04-02.148638.json

@@ -0,0 +1,337 @@
+{
+  "results": {
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 1319,
+      "exact_match,strict-match": 0.0075815011372251705,
+      "exact_match_stderr,strict-match": 0.002389281512077213,
+      "exact_match,flexible-extract": 0.3775587566338135,
+      "exact_match_stderr,flexible-extract": 0.013353150666358537
+    },
+    "minerva_math500": {
+      "name": "minerva_math500",
+      "alias": "minerva_math500",
+      "sample_len": 500,
+      "exact_match,none": 0.11,
+      "exact_match_stderr,none": 0.01400686919941566,
+      "math_verify,none": 0.164,
+      "math_verify_stderr,none": 0.016575811142446696
+    },
+    "ifeval": {
+      "name": "ifeval",
+      "alias": "ifeval",
+      "sample_len": 541,
+      "prompt_level_strict_acc,none": 0.5508317929759704,
+      "prompt_level_strict_acc_stderr,none": 0.021405093233588298,
+      "inst_level_strict_acc,none": 0.6714628297362111,
+      "inst_level_strict_acc_stderr,none": "N/A",
+      "prompt_level_loose_acc,none": 0.5804066543438078,
+      "prompt_level_loose_acc_stderr,none": 0.021236532548855144,
+      "inst_level_loose_acc,none": 0.7014388489208633,
+      "inst_level_loose_acc_stderr,none": "N/A"
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    },
+    "ifeval": {
+      "task": "ifeval",
+      "dataset_path": "google/IFEval",
+      "test_split": "train",
+      "doc_to_text": "prompt",
+      "doc_to_target": 0,
+      "unsafe_code": false,
+      "process_results": "def process_results(doc, results):\n    inp = InputExample(\n        key=doc[\"key\"],\n        instruction_id_list=doc[\"instruction_id_list\"],\n        prompt=doc[\"prompt\"],\n        kwargs=doc[\"kwargs\"],\n    )\n    response = results[0]\n\n    out_strict = test_instruction_following_strict(inp, response)\n    out_loose = test_instruction_following_loose(inp, response)\n\n    return {\n        \"prompt_level_strict_acc\": out_strict.follow_all_instructions,\n        \"inst_level_strict_acc\": out_strict.follow_instruction_list,\n        \"prompt_level_loose_acc\": out_loose.follow_all_instructions,\n        \"inst_level_loose_acc\": out_loose.follow_instruction_list,\n    }\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "prompt",
+        "doc_to_choice": null,
+        "doc_to_target": 0,
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "prompt_level_strict_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_strict_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        },
+        {
+          "metric": "prompt_level_loose_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_loose_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 4.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/ifeval/ifeval.yaml"
+      }
+    },
+    "minerva_math500": {
+      "task": "minerva_math500",
+      "dataset_path": "HuggingFaceH4/MATH-500",
+      "dataset_name": "default",
+      "test_split": "test",
+      "process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n    def _process_doc(doc: dict) -> dict:\n        out_doc = {\n            \"problem\": doc[\"problem\"],\n            \"solution\": doc[\"solution\"],\n            \"answer\": normalize_final_answer(\n                remove_boxed(last_boxed_only_string(doc[\"solution\"]))\n            ),\n        }\n        if getattr(doc, \"few_shot\", None) is not None:\n            out_doc[\"few_shot\"] = True\n        return out_doc\n\n    return dataset.map(_process_doc)\n",
+      "doc_to_text": "def doc_to_text(doc: dict) -> str:\n    return \"Problem:\" + \"\\n\" + doc[\"problem\"] + \"\\n\\n\" + \"Solution:\"\n",
+      "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+      "unsafe_code": false,
+      "process_results": "def process_results(doc: dict, results: list[str]) -> dict[str, int]:\n    candidates = results[0]\n\n    unnormalized_answer = get_unnormalized_answer(candidates)\n    answer = normalize_final_answer(unnormalized_answer)\n\n    if is_equiv(answer, doc[\"answer\"]):\n        retval = 1\n    else:\n        retval = 0\n\n    # math_verify\n    _mvres = verify(\n        gold=parse(doc[\"solution\"]),\n        target=parse(candidates),\n    )\n    mathval = 1 if _mvres else 0\n\n    res = {\n        \"exact_match\": retval,\n        \"math_verify\": mathval,\n    }\n    return res\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": "<function process_docs at 0x7d91dd921b20>",
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x7d91dd9239c0>",
+        "doc_to_text": "<function doc_to_text at 0x7d91df0e98a0>",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 4,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "math_verify",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Problem:"
+        ],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/minerva_math/minerva_math500.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k_cot_zeroshot": 3.0,
+    "ifeval": 4.0,
+    "minerva_math500": 3.0
+  },
+  "n-shot": {
+    "gsm8k_cot_zeroshot": 0,
+    "ifeval": 0,
+    "minerva_math500": 4
+  },
+  "higher_is_better": {
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    },
+    "ifeval": {
+      "prompt_level_strict_acc": true,
+      "inst_level_strict_acc": true,
+      "prompt_level_loose_acc": true,
+      "inst_level_loose_acc": true
+    },
+    "minerva_math500": {
+      "exact_match": true,
+      "math_verify": true
+    }
+  },
+  "n-samples": {
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 1319
+    },
+    "minerva_math500": {
+      "original": 500,
+      "effective": 500
+    },
+    "ifeval": {
+      "original": 541,
+      "effective": 541
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {
+      "max_gen_toks": 1280
+    },
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "4cd474a",
+  "date": 1784527132.8290846,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.11\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k_cot_zeroshot": "77ad80425a4cb85562b27f164a9c92977f775bb66a0ec4d374e4ee7b4656a7f5",
+    "minerva_math500": "5784cb060f783763ec8f8e234eef61589d348055c48e8adc0073445383258f85",
+    "ifeval": "bdcf10fc42fb3b39fd60af9d347a998b5bcb0a6782900c760b7dcc67ef0109d7"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "315.685363393015"
+}

evals/general_suite/healed/reap_keep50_s1224_long500_step500/student/results_2026-07-19T23-07-03.090538.json

@@ -0,0 +1,225 @@
+{
+  "results": {
+    "humaneval": {
+      "name": "humaneval",
+      "alias": "humaneval",
+      "sample_len": 164,
+      "pass@1,create_test": 0.2804878048780488,
+      "pass@1_stderr,create_test": 0.03518700228801581
+    },
+    "mbpp": {
+      "name": "mbpp",
+      "alias": "mbpp",
+      "sample_len": 500,
+      "pass_at_1,none": 0.236,
+      "pass_at_1_stderr,none": 0.019008699622084728
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "humaneval": {
+      "task": "humaneval",
+      "dataset_path": "openai/openai_humaneval",
+      "test_split": "test",
+      "doc_to_text": "{{prompt}}",
+      "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "{{prompt}}",
+        "doc_to_choice": null,
+        "doc_to_target": "{{test}}\ncheck({{entry_point}})",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "def pass_at_k(references: list[str], predictions: list[list[str]], k: list[int] = None):\n    global compute_\n    assert k is not None\n    if isinstance(k, int):\n        k = [k]\n    res = compute_.compute(\n        references=references,\n        predictions=predictions,\n        k=k,\n    )\n    return res[0]\n",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "k": [
+            1
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "\nclass",
+          "\ndef",
+          "\n#",
+          "\nif",
+          "\nprint"
+        ],
+        "max_gen_toks": 1024,
+        "do_sample": false
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "create_test",
+          "filter": [
+            {
+              "function": "custom",
+              "filter_fn": "<function build_predictions at 0x72bbe4e6eca0>"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/completions",
+        "tokenizer": "outputs/healed/grid_general_fairness/reap_keep50_s1224_long500/step0500",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/humaneval/humaneval.yaml"
+      }
+    },
+    "mbpp": {
+      "task": "mbpp",
+      "dataset_path": "google-research-datasets/mbpp",
+      "dataset_name": "full",
+      "test_split": "test",
+      "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+      "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+      "unsafe_code": true,
+      "description": "",
+      "target_delimiter": "",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x72bcadf079c0>",
+        "doc_to_text": "You are an expert Python programmer, and here is your task: {{text}} Your code should pass these tests:\n\n{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}\n[BEGIN]\n",
+        "doc_to_choice": null,
+        "doc_to_target": "{% if is_fewshot is defined %}{{code}}\n[DONE]{% else %}{{test_list[0]}}\n{{test_list[1]}}\n{{test_list[2]}}{% endif %}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": ""
+      },
+      "num_fewshot": 3,
+      "metric_list": [
+        {
+          "metric": "def pass_at_1(\n    references: Union[str, list[str]], predictions: Union[str, list[list[str]]]\n) -> float:\n    if isinstance(references, str):\n        references = [references]\n    if isinstance(predictions[0], str):\n        predictions = [[p] for p in predictions]\n    return pass_at_k.compute(\n        references=references,\n        predictions=predictions,\n        k=[1],\n    )[0][\"pass@1\"]\n",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "[DONE]"
+        ],
+        "do_sample": false
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 1.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/completions",
+        "tokenizer": "outputs/healed/grid_general_fairness/reap_keep50_s1224_long500/step0500",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/mbpp/mbpp.yaml"
+      }
+    }
+  },
+  "versions": {
+    "humaneval": 1.0,
+    "mbpp": 1.0
+  },
+  "n-shot": {
+    "humaneval": 0,
+    "mbpp": 3
+  },
+  "higher_is_better": {
+    "humaneval": {
+      "pass_at_k": true,
+      "pass@1": true
+    },
+    "mbpp": {
+      "pass_at_1": true
+    }
+  },
+  "n-samples": {
+    "humaneval": {
+      "original": 164,
+      "effective": 164
+    },
+    "mbpp": {
+      "original": 500,
+      "effective": 500
+    }
+  },
+  "config": {
+    "model": "local-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/completions",
+      "tokenizer": "outputs/healed/grid_general_fairness/reap_keep50_s1224_long500/step0500",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {},
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "4cd474a",
+  "date": 1784527451.69169,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.11\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "tokenizer_pad_token": [
+    "<pad>",
+    "50280"
+  ],
+  "tokenizer_eos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "tokenizer_bos_token": [
+    "|||IP_ADDRESS|||",
+    "50279"
+  ],
+  "eot_token_id": 50279,
+  "max_length": 2047,
+  "task_hashes": {
+    "humaneval": "c122632f0bdffdd0162cc8033a879dfe9d3b35e82c52eda6a767069e0a2beb40",
+    "mbpp": "62356f643ed511c895814c093ddb55451941156ba9febb31dc87c6e00d9723e3"
+  },
+  "model_source": "local-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": null,
+  "chat_template": null,
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "177.78676151300897"
+}

evals/general_suite/healed/uniform_keep25_s1225/student/results_2026-07-17T19-15-00.481091.json

@@ -0,0 +1,337 @@
+{
+  "results": {
+    "gsm8k_cot_zeroshot": {
+      "name": "gsm8k_cot_zeroshot",
+      "alias": "gsm8k_cot_zeroshot",
+      "sample_len": 1319,
+      "exact_match,strict-match": 0.000758150113722517,
+      "exact_match_stderr,strict-match": 0.0007581501137225197,
+      "exact_match,flexible-extract": 0.16679302501895377,
+      "exact_match_stderr,flexible-extract": 0.010268516042629511
+    },
+    "minerva_math500": {
+      "name": "minerva_math500",
+      "alias": "minerva_math500",
+      "sample_len": 500,
+      "exact_match,none": 0.038,
+      "exact_match_stderr,none": 0.008559118940292168,
+      "math_verify,none": 0.064,
+      "math_verify_stderr,none": 0.01095664621097098
+    },
+    "ifeval": {
+      "name": "ifeval",
+      "alias": "ifeval",
+      "sample_len": 541,
+      "prompt_level_strict_acc,none": 0.43253234750462105,
+      "prompt_level_strict_acc_stderr,none": 0.021319792398178095,
+      "inst_level_strict_acc,none": 0.5695443645083933,
+      "inst_level_strict_acc_stderr,none": "N/A",
+      "prompt_level_loose_acc,none": 0.4676524953789279,
+      "prompt_level_loose_acc_stderr,none": 0.02147149872051707,
+      "inst_level_loose_acc,none": 0.5935251798561151,
+      "inst_level_loose_acc_stderr,none": "N/A"
+    }
+  },
+  "group_subtasks": {},
+  "configs": {
+    "gsm8k_cot_zeroshot": {
+      "task": "gsm8k_cot_zeroshot",
+      "dataset_path": "openai/gsm8k",
+      "dataset_name": "main",
+      "training_split": "train",
+      "test_split": "test",
+      "fewshot_split": "train",
+      "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+      "doc_to_target": "{{answer}}",
+      "unsafe_code": false,
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": "train",
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "Q: {{question}}\nA: Let's think step by step.",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true,
+          "ignore_case": true,
+          "ignore_punctuation": false,
+          "regexes_to_ignore": [
+            ",",
+            "\\$",
+            "(?s).*#### ",
+            "\\.$"
+          ]
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Q:",
+          "</s>",
+          "<|im_end|>"
+        ],
+        "do_sample": false,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "filter_list": [
+        {
+          "name": "strict-match",
+          "filter": [
+            {
+              "function": "regex",
+              "regex_pattern": "The answer is (\\-?[0-9\\.\\,]+)."
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        },
+        {
+          "name": "flexible-extract",
+          "filter": [
+            {
+              "function": "regex",
+              "group_select": -1,
+              "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)"
+            },
+            {
+              "function": "take_first"
+            }
+          ]
+        }
+      ],
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/gsm8k/gsm8k-cot-zeroshot.yaml"
+      }
+    },
+    "ifeval": {
+      "task": "ifeval",
+      "dataset_path": "google/IFEval",
+      "test_split": "train",
+      "doc_to_text": "prompt",
+      "doc_to_target": 0,
+      "unsafe_code": false,
+      "process_results": "def process_results(doc, results):\n    inp = InputExample(\n        key=doc[\"key\"],\n        instruction_id_list=doc[\"instruction_id_list\"],\n        prompt=doc[\"prompt\"],\n        kwargs=doc[\"kwargs\"],\n    )\n    response = results[0]\n\n    out_strict = test_instruction_following_strict(inp, response)\n    out_loose = test_instruction_following_loose(inp, response)\n\n    return {\n        \"prompt_level_strict_acc\": out_strict.follow_all_instructions,\n        \"inst_level_strict_acc\": out_strict.follow_instruction_list,\n        \"prompt_level_loose_acc\": out_loose.follow_all_instructions,\n        \"inst_level_loose_acc\": out_loose.follow_instruction_list,\n    }\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "default",
+        "split": null,
+        "process_docs": null,
+        "fewshot_indices": null,
+        "samples": null,
+        "doc_to_text": "prompt",
+        "doc_to_choice": null,
+        "doc_to_target": 0,
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 0,
+      "metric_list": [
+        {
+          "metric": "prompt_level_strict_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_strict_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        },
+        {
+          "metric": "prompt_level_loose_acc",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "inst_level_loose_acc",
+          "aggregation": "def agg_inst_level_acc(items):\n    flat_items = [item for sublist in items for item in sublist]\n    inst_level_acc = sum(flat_items) / len(flat_items)\n    return inst_level_acc\n",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 4.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/ifeval/ifeval.yaml"
+      }
+    },
+    "minerva_math500": {
+      "task": "minerva_math500",
+      "dataset_path": "HuggingFaceH4/MATH-500",
+      "dataset_name": "default",
+      "test_split": "test",
+      "process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n    def _process_doc(doc: dict) -> dict:\n        out_doc = {\n            \"problem\": doc[\"problem\"],\n            \"solution\": doc[\"solution\"],\n            \"answer\": normalize_final_answer(\n                remove_boxed(last_boxed_only_string(doc[\"solution\"]))\n            ),\n        }\n        if getattr(doc, \"few_shot\", None) is not None:\n            out_doc[\"few_shot\"] = True\n        return out_doc\n\n    return dataset.map(_process_doc)\n",
+      "doc_to_text": "def doc_to_text(doc: dict) -> str:\n    return \"Problem:\" + \"\\n\" + doc[\"problem\"] + \"\\n\\n\" + \"Solution:\"\n",
+      "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+      "unsafe_code": false,
+      "process_results": "def process_results(doc: dict, results: list[str]) -> dict[str, int]:\n    candidates = results[0]\n\n    unnormalized_answer = get_unnormalized_answer(candidates)\n    answer = normalize_final_answer(unnormalized_answer)\n\n    if is_equiv(answer, doc[\"answer\"]):\n        retval = 1\n    else:\n        retval = 0\n\n    # math_verify\n    _mvres = verify(\n        gold=parse(doc[\"solution\"]),\n        target=parse(candidates),\n    )\n    mathval = 1 if _mvres else 0\n\n    res = {\n        \"exact_match\": retval,\n        \"math_verify\": mathval,\n    }\n    return res\n",
+      "description": "",
+      "target_delimiter": " ",
+      "fewshot_delimiter": "\n\n",
+      "fewshot_config": {
+        "sampler": "first_n",
+        "split": null,
+        "process_docs": "<function process_docs at 0x764598c2db20>",
+        "fewshot_indices": null,
+        "samples": "<function list_fewshot_samples at 0x764598c2f9c0>",
+        "doc_to_text": "<function doc_to_text at 0x76459a3f58a0>",
+        "doc_to_choice": null,
+        "doc_to_target": "{{answer if few_shot is undefined else solution}}",
+        "gen_prefix": null,
+        "fewshot_delimiter": "\n\n",
+        "target_delimiter": " "
+      },
+      "num_fewshot": 4,
+      "metric_list": [
+        {
+          "metric": "exact_match",
+          "aggregation": "mean",
+          "higher_is_better": true
+        },
+        {
+          "metric": "math_verify",
+          "aggregation": "mean",
+          "higher_is_better": true
+        }
+      ],
+      "output_type": "generate_until",
+      "generation_kwargs": {
+        "until": [
+          "Problem:"
+        ],
+        "do_sample": false,
+        "temperature": 0.0,
+        "max_gen_toks": 1280
+      },
+      "repeats": 1,
+      "should_decontaminate": false,
+      "metadata": {
+        "version": 3.0,
+        "model": "student",
+        "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+        "num_concurrent": 48,
+        "tokenized_requests": false,
+        "max_retries": 3,
+        "config_source": "/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/lm_eval/tasks/minerva_math/minerva_math500.yaml"
+      }
+    }
+  },
+  "versions": {
+    "gsm8k_cot_zeroshot": 3.0,
+    "ifeval": 4.0,
+    "minerva_math500": 3.0
+  },
+  "n-shot": {
+    "gsm8k_cot_zeroshot": 0,
+    "ifeval": 0,
+    "minerva_math500": 4
+  },
+  "higher_is_better": {
+    "gsm8k_cot_zeroshot": {
+      "exact_match": true
+    },
+    "ifeval": {
+      "prompt_level_strict_acc": true,
+      "inst_level_strict_acc": true,
+      "prompt_level_loose_acc": true,
+      "inst_level_loose_acc": true
+    },
+    "minerva_math500": {
+      "exact_match": true,
+      "math_verify": true
+    }
+  },
+  "n-samples": {
+    "gsm8k_cot_zeroshot": {
+      "original": 1319,
+      "effective": 1319
+    },
+    "minerva_math500": {
+      "original": 500,
+      "effective": 500
+    },
+    "ifeval": {
+      "original": 541,
+      "effective": 541
+    }
+  },
+  "config": {
+    "model": "local-chat-completions",
+    "model_args": {
+      "model": "student",
+      "base_url": "http://127.0.0.1:8421/v1/chat/completions",
+      "num_concurrent": 48,
+      "tokenized_requests": false,
+      "max_retries": 3
+    },
+    "batch_size": 1,
+    "batch_sizes": [],
+    "device": "cuda:0",
+    "use_cache": null,
+    "limit": null,
+    "bootstrap_iters": 100000,
+    "gen_kwargs": {
+      "max_gen_toks": 1280
+    },
+    "random_seed": 0,
+    "numpy_seed": 1234,
+    "torch_seed": 1234,
+    "fewshot_seed": 1234
+  },
+  "git_hash": "247c7f0",
+  "date": 1784340269.0403202,
+  "pretty_env_info": "PyTorch version: 2.12.1+cu126\nIs debug build: False\nCUDA used to build PyTorch: 12.6\nROCM used to build PyTorch: N/A\n\nOS: Pop!_OS 22.04 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.12.12 (main, Oct 31 2025, 23:02:31) [Clang 21.1.4 ] (64-bit runtime)\nPython platform: Linux-7.0.11-76070011-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.8.93\nCUDA_MODULE_LOADING set to: \nGPU models and configuration: \nGPU 0: NVIDIA GeForce RTX 3090\nGPU 1: NVIDIA GeForce RTX 3090\nGPU 2: NVIDIA GeForce RTX 3090\n\nNvidia driver version: 580.159.03\ncuDNN version: Could not collect\nIs XPU available: False\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\nCaching allocator config: N/A\n\nCPU:\nArchitecture:                            x86_64\nCPU op-mode(s):                          32-bit, 64-bit\nAddress sizes:                           43 bits physical, 48 bits virtual\nByte Order:                              Little Endian\nCPU(s):                                  48\nOn-line CPU(s) list:                     0-47\nVendor ID:                               AuthenticAMD\nModel name:                              AMD EPYC 7402P 24-Core Processor\nCPU family:                              23\nModel:                                   49\nThread(s) per core:                      2\nCore(s) per socket:                      24\nSocket(s):                               1\nStepping:                                0\nFrequency boost:                         enabled\nCPU max MHz:                             3350.8049\nCPU min MHz:                             1500.0000\nBogoMIPS:                                5600.23\nFlags:                                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sev sev_es\nVirtualization:                          AMD-V\nL1d cache:                               768 KiB (24 instances)\nL1i cache:                               768 KiB (24 instances)\nL2 cache:                                12 MiB (24 instances)\nL3 cache:                                128 MiB (8 instances)\nNUMA node(s):                            1\nNUMA node0 CPU(s):                       0-47\nVulnerability Gather data sampling:      Not affected\nVulnerability Ghostwrite:                Not affected\nVulnerability Indirect target selection: Not affected\nVulnerability Itlb multihit:             Not affected\nVulnerability L1tf:                      Not affected\nVulnerability Mds:                       Not affected\nVulnerability Meltdown:                  Not affected\nVulnerability Mmio stale data:           Not affected\nVulnerability Old microcode:             Not affected\nVulnerability Reg file data sampling:    Not affected\nVulnerability Retbleed:                  Mitigation; untrained return thunk; SMT enabled with STIBP protection\nVulnerability Spec rstack overflow:      Mitigation; Safe RET\nVulnerability Spec store bypass:         Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1:                Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2:                Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds:                     Not affected\nVulnerability Tsa:                       Not affected\nVulnerability Tsx async abort:           Not affected\nVulnerability Vmscape:                   Mitigation; IBPB before exit to userspace\n\nVersions of relevant libraries:\n[pip3] Could not collect\n[conda] Could not collect",
+  "transformers_version": "4.57.6",
+  "lm_eval_version": "0.4.12",
+  "upper_git_hash": null,
+  "task_hashes": {
+    "gsm8k_cot_zeroshot": "77ad80425a4cb85562b27f164a9c92977f775bb66a0ec4d374e4ee7b4656a7f5",
+    "minerva_math500": "5784cb060f783763ec8f8e234eef61589d348055c48e8adc0073445383258f85",
+    "ifeval": "bdcf10fc42fb3b39fd60af9d347a998b5bcb0a6782900c760b7dcc67ef0109d7"
+  },
+  "model_source": "local-chat-completions",
+  "model_name": "student",
+  "model_name_sanitized": "student",
+  "system_instruction": null,
+  "system_instruction_sha": null,
+  "fewshot_as_multiturn": true,
+  "chat_template": "",
+  "chat_template_sha": null,
+  "total_evaluation_time_seconds": "637.8067643560062"
+}