| [ |
| { |
| "title": "Accuracy", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - EM", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nQuasi-exact match: Fraction of instances that the predicted output matches a correct reference up to light processing.", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": { |
| "metric": "EM", |
| "run_group": "CivilComments" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "value": 0.0, |
| "description": "min=0, mean=0, max=0, sum=0 (18)", |
| "style": { |
| "font-weight": "bold" |
| }, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_accuracy.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_accuracy.json" |
| } |
| ], |
| "name": "accuracy" |
| }, |
| { |
| "title": "Calibration", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - ECE (10-bin)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\n10-bin expected calibration error: The average difference between the model's confidence and accuracy, averaged across 10 bins where each bin contains an equal number of points (only computed for classification tasks). Warning - not reliable for small datasets (e.g., with < 300 examples) because each bin will have very few examples.", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "ECE (10-bin)", |
| "run_group": "CivilComments" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "value": 3.3724591072781132e-12, |
| "description": "min=0.0, mean=0.0, max=0.0, sum=0.0 (18)", |
| "style": { |
| "font-weight": "bold" |
| }, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_calibration.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_calibration.json" |
| } |
| ], |
| "name": "calibration" |
| }, |
| { |
| "title": "Robustness", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - EM (Robustness)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nQuasi-exact match: Fraction of instances that the predicted output matches a correct reference up to light processing.\n- Perturbation Robustness: Computes worst case over different robustness perturbations (misspellings, formatting, contrast sets).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": { |
| "metric": "EM", |
| "run_group": "CivilComments", |
| "perturbation": "Robustness" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "value": 0.0, |
| "description": "min=0, mean=0, max=0, sum=0 (18)", |
| "style": { |
| "font-weight": "bold" |
| }, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_robustness.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_robustness.json" |
| } |
| ], |
| "name": "robustness" |
| }, |
| { |
| "title": "Fairness", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - EM (Fairness)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nQuasi-exact match: Fraction of instances that the predicted output matches a correct reference up to light processing.\n- Perturbation Fairness: Computes worst case over different fairness perturbations (changing dialect, race of names, gender).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": { |
| "metric": "EM", |
| "run_group": "CivilComments", |
| "perturbation": "Fairness" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "value": 0.0, |
| "description": "min=0, mean=0, max=0, sum=0 (18)", |
| "style": { |
| "font-weight": "bold" |
| }, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_fairness.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_fairness.json" |
| } |
| ], |
| "name": "fairness" |
| }, |
| { |
| "title": "Bias", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - Stereotypes (race)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nStereotypical associations (race, profession): Measures uneven association of racial groups (Asian, Hispanic, White) with target professions. This measurement is based on cooccurence statistics between the racially-associated names (word list from [Garg et al., 2018](https://www.pnas.org/doi/10.1073/pnas.1720347115); race associations based on US Census statistics) and the target professions (word list from [Bolukbasi et al., 2016](https://papers.nips.cc/paper/2016/hash/a486cd07e4ac3d270571622f4f316ec5-Abstract.html)).", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "Stereotypes (race)", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - Stereotypes (gender)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nStereotypical associations (gender, profession): Measures uneven association of gender groups (male, female) with target professions. This measurement is based on cooccurence statistics between the gender terms (word list from [Bolukbasi et al., 2016](https://papers.nips.cc/paper/2016/hash/a486cd07e4ac3d270571622f4f316ec5-Abstract.html)) and the target professions (word list from [Bolukbasi et al., 2016](https://papers.nips.cc/paper/2016/hash/a486cd07e4ac3d270571622f4f316ec5-Abstract.html)).", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "Stereotypes (gender)", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - Representation (race)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nDemographic representation (race): Measures uneven representation of racial groups (Asian, Hispanic, White). This measurement is based on disparities in the frequency statistics across racially-associated names (word list from [Garg et al., 2018](https://www.pnas.org/doi/10.1073/pnas.1720347115); race associations based on US Census statistics).", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "Representation (race)", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - Representation (gender)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nDemographic representation (gender): Measures uneven representation of gender groups (male, female). This measurement is based on disparities in the frequency statistics across gender terms (word list from [Bolukbasi et al., 2016](https://papers.nips.cc/paper/2016/hash/a486cd07e4ac3d270571622f4f316ec5-Abstract.html)).", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "Representation (gender)", |
| "run_group": "CivilComments" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "description": "(0)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "description": "(0)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "description": "(0)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "description": "(0)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_bias.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_bias.json" |
| } |
| ], |
| "name": "bias" |
| }, |
| { |
| "title": "Toxicity", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - Toxic fraction", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nToxic fraction: Fraction of model outputs that are toxic (based on the PerspectiveAPI toxicity classifier).", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "Toxic fraction", |
| "run_group": "CivilComments" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "description": "9 matching runs, but no matching metrics", |
| "markdown": false |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_toxicity.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_toxicity.json" |
| } |
| ], |
| "name": "toxicity" |
| }, |
| { |
| "title": "Efficiency", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - Denoised inference time (s)", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\nDenoised inference runtime (s): Average time to process a request to the model minus performance contention by using profiled runtimes from multiple trials of SyntheticEfficiencyScenario.", |
| "markdown": false, |
| "lower_is_better": true, |
| "metadata": { |
| "metric": "Denoised inference time (s)", |
| "run_group": "CivilComments" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "description": "9 matching runs, but no matching metrics", |
| "markdown": false |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_efficiency.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_efficiency.json" |
| } |
| ], |
| "name": "efficiency" |
| }, |
| { |
| "title": "General information", |
| "header": [ |
| { |
| "value": "Model", |
| "markdown": false, |
| "metadata": {} |
| }, |
| { |
| "value": "Mean win rate", |
| "description": "How many models this model outperforms on average (over columns).", |
| "markdown": false, |
| "lower_is_better": false, |
| "metadata": {} |
| }, |
| { |
| "value": "CivilComments - # eval", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\n# eval: Number of evaluation instances.", |
| "markdown": false, |
| "metadata": { |
| "metric": "# eval", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - # train", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\n# train: Number of training instances (e.g., in-context examples).", |
| "markdown": false, |
| "metadata": { |
| "metric": "# train", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - truncated", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\ntruncated: Fraction of instances where the prompt itself was truncated (implies that there were no in-context examples).", |
| "markdown": false, |
| "metadata": { |
| "metric": "truncated", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - # prompt tokens", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\n# prompt tokens: Number of tokens in the prompt.", |
| "markdown": false, |
| "metadata": { |
| "metric": "# prompt tokens", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - # output tokens", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\n# output tokens: Actual number of output tokens.", |
| "markdown": false, |
| "metadata": { |
| "metric": "# output tokens", |
| "run_group": "CivilComments" |
| } |
| }, |
| { |
| "value": "CivilComments - # trials", |
| "description": "The CivilComments benchmark for toxicity detection [(Borkan et al., 2019)](https://arxiv.org/pdf/1903.04561.pdf).\n\n# trials: Number of trials, where in each trial we choose an independent, random set of training instances.", |
| "markdown": false, |
| "metadata": { |
| "metric": "# trials", |
| "run_group": "CivilComments" |
| } |
| } |
| ], |
| "rows": [ |
| [ |
| { |
| "value": "EleutherAI/pythia-2.8b", |
| "description": "", |
| "markdown": false |
| }, |
| { |
| "markdown": false |
| }, |
| { |
| "value": 371.55555555555554, |
| "description": "min=74, mean=371.556, max=683, sum=6688 (18)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "value": 5.0, |
| "description": "min=5, mean=5, max=5, sum=90 (18)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "value": 0.0, |
| "description": "min=0, mean=0, max=0, sum=0 (18)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "value": 771.6539847352628, |
| "description": "min=360.976, mean=771.654, max=1282.4, sum=13889.772 (18)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "value": 4.999712147380541, |
| "description": "min=4.995, mean=5.0, max=5, sum=89.995 (18)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| }, |
| { |
| "value": 1.0, |
| "description": "min=1, mean=1, max=1, sum=18 (18)", |
| "style": {}, |
| "markdown": false, |
| "run_spec_names": [ |
| "civil_comments:demographic=LGBTQ,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=all,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=black,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=christian,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=female,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=male,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=muslim,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=other_religions,model=EleutherAI_pythia-2.8b,data_augmentation=canonical", |
| "civil_comments:demographic=white,model=EleutherAI_pythia-2.8b,data_augmentation=canonical" |
| ] |
| } |
| ] |
| ], |
| "links": [ |
| { |
| "text": "LaTeX", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/latex/toxicity_detection_general_information.tex" |
| }, |
| { |
| "text": "JSON", |
| "href": "benchmark_output/runs/classic_pythia-2.8b-step2/groups/json/toxicity_detection_general_information.json" |
| } |
| ], |
| "name": "general_information" |
| } |
| ] |
