Merge branch 'main' into ai2d

lm_eval/tasks/global_mmlu/default/zh/_zh_template_yaml

+dataset_path: CohereForAI/Global-MMLU-Lite
+dataset_name: zh
+test_split: test
+fewshot_split: dev
+fewshot_config:
+  sampler: default
+output_type: multiple_choice
+doc_to_text: "{{question.strip()}}\nA. {{option_a}}\nB. {{option_b}}\nC. {{option_c}}\nD. {{option_d}}\nAnswer:"
+doc_to_choice: ["A", "B", "C", "D"]
+doc_to_target: answer
+metric_list:
+  - metric: acc
+    aggregation: mean
+    higher_is_better: true
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/default/zh/global_mmlu_zh_business.yaml

+# Generated by _generate_configs.py
+include: _zh_template_yaml
+process_docs: !function utils.process_business
+task: global_mmlu_zh_business

lm_eval/tasks/global_mmlu/default/zh/global_mmlu_zh_humanities.yaml

+# Generated by _generate_configs.py
+include: _zh_template_yaml
+process_docs: !function utils.process_humanities
+task: global_mmlu_zh_humanities

lm_eval/tasks/global_mmlu/default/zh/global_mmlu_zh_medical.yaml

+# Generated by _generate_configs.py
+include: _zh_template_yaml
+process_docs: !function utils.process_medical
+task: global_mmlu_zh_medical

lm_eval/tasks/global_mmlu/default/zh/global_mmlu_zh_other.yaml

+# Generated by _generate_configs.py
+include: _zh_template_yaml
+process_docs: !function utils.process_other
+task: global_mmlu_zh_other

lm_eval/tasks/global_mmlu/default/zh/global_mmlu_zh_social_sciences.yaml

+# Generated by _generate_configs.py
+include: _zh_template_yaml
+process_docs: !function utils.process_social_sciences
+task: global_mmlu_zh_social_sciences

lm_eval/tasks/global_mmlu/default/zh/global_mmlu_zh_stem.yaml

+# Generated by _generate_configs.py
+include: _zh_template_yaml
+process_docs: !function utils.process_stem
+task: global_mmlu_zh_stem

lm_eval/tasks/global_mmlu/default/zh/utils.py

+from functools import partial
+
+
+CATEGORIES = ["Business", "Humanities", "Medical", "Other", "STEM", "Social Sciences"]
+
+
+def process_docs(dataset, category):
+    return dataset.filter(lambda x: x["subject_category"] == category)
+
+
+process_functions = {
+    f"process_{category.lower().replace(' ', '_')}": partial(
+        process_docs, category=category
+    )
+    for category in CATEGORIES
+}
+
+globals().update(process_functions)

lm_eval/tasks/global_mmlu/full/am/_am_template_yaml

+dataset_path: CohereForAI/Global-MMLU
+dataset_name: am
+test_split: test
+fewshot_split: dev
+fewshot_config:
+  sampler: first_n
+output_type: multiple_choice
+doc_to_text: "{{question.strip()}}\nA. {{option_a}}\nB. {{option_b}}\nC. {{option_c}}\nD. {{option_d}}\nAnswer:"
+doc_to_choice: ["A", "B", "C", "D"]
+doc_to_target: answer
+metric_list:
+  - metric: acc
+    aggregation: mean
+    higher_is_better: true
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/full/am/_global_mmlu_full_am.yaml

+group: global_mmlu_full_am
+task:
+  - global_mmlu_full_am_stem
+  - global_mmlu_full_am_other
+  - global_mmlu_full_am_social_sciences
+  - global_mmlu_full_am_humanities
+aggregate_metric_list:
+  - metric: acc
+    weight_by_size: True
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/full/am/_global_mmlu_full_am_humanities.yaml

+group: global_mmlu_full_am_humanities
+task:
+  - global_mmlu_full_am_humanities_tasks
+aggregate_metric_list:
+  - metric: acc
+    weight_by_size: true
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/full/am/_global_mmlu_full_am_other.yaml

+group: global_mmlu_full_am_other
+task:
+  - global_mmlu_full_am_other_tasks
+aggregate_metric_list:
+  - metric: acc
+    weight_by_size: true
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/full/am/_global_mmlu_full_am_social_sciences.yaml

+group: global_mmlu_full_am_social_sciences
+task:
+  - global_mmlu_full_am_social_sciences_tasks
+aggregate_metric_list:
+  - metric: acc
+    weight_by_size: true
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/full/am/_global_mmlu_full_am_stem.yaml

+group: global_mmlu_full_am_stem
+task:
+  - global_mmlu_full_am_stem_tasks
+aggregate_metric_list:
+  - metric: acc
+    weight_by_size: true
+metadata:
+  version: 0.0

lm_eval/tasks/global_mmlu/full/am/global_mmlu_full_am_abstract_algebra.yaml

+# Generated by _generate_configs.py
+include: _am_template_yaml
+process_docs: !function utils.process_abstract_algebra
+tag: global_mmlu_full_am_stem_tasks
+task: global_mmlu_full_am_abstract_algebra

lm_eval/tasks/global_mmlu/full/am/global_mmlu_full_am_anatomy.yaml

+# Generated by _generate_configs.py
+include: _am_template_yaml
+process_docs: !function utils.process_anatomy
+tag: global_mmlu_full_am_stem_tasks
+task: global_mmlu_full_am_anatomy

lm_eval/tasks/global_mmlu/full/am/global_mmlu_full_am_astronomy.yaml

+# Generated by _generate_configs.py
+include: _am_template_yaml
+process_docs: !function utils.process_astronomy
+tag: global_mmlu_full_am_stem_tasks
+task: global_mmlu_full_am_astronomy

lm_eval/tasks/global_mmlu/full/am/global_mmlu_full_am_business_ethics.yaml

+# Generated by _generate_configs.py
+include: _am_template_yaml
+process_docs: !function utils.process_business_ethics
+tag: global_mmlu_full_am_other_tasks
+task: global_mmlu_full_am_business_ethics

lm_eval/tasks/global_mmlu/full/am/global_mmlu_full_am_clinical_knowledge.yaml

+# Generated by _generate_configs.py
+include: _am_template_yaml
+process_docs: !function utils.process_clinical_knowledge
+tag: global_mmlu_full_am_other_tasks
+task: global_mmlu_full_am_clinical_knowledge

lm_eval/tasks/global_mmlu/full/am/global_mmlu_full_am_college_biology.yaml

+# Generated by _generate_configs.py
+include: _am_template_yaml
+process_docs: !function utils.process_college_biology
+tag: global_mmlu_full_am_stem_tasks
+task: global_mmlu_full_am_college_biology