diff --git "a/evaluations/ar/Llama-3.3-70B-Instruct/arabicmmlu_0_shot.json" "b/evaluations/ar/Llama-3.3-70B-Instruct/arabicmmlu_0_shot.json" new file mode 100644--- /dev/null +++ "b/evaluations/ar/Llama-3.3-70B-Instruct/arabicmmlu_0_shot.json" @@ -0,0 +1,2051 @@ +{ + "results": { + "arabicmmlu": { + "acc,none": 0.7200968523002421, + "acc_stderr,none": 0.003653809830387355, + "alias": "arabicmmlu" + }, + "arabicmmlu_humanities": { + "acc,none": 0.7367695700110254, + "acc_stderr,none": 0.007118478408616655, + "alias": " - Humanities" + }, + "arabicmmlu_high_history": { + "alias": " - High History", + "acc,none": 0.5644736842105263, + "acc_stderr,none": 0.01799733343022178 + }, + "arabicmmlu_high_islamic_studies": { + "alias": " - High Islamic Studies", + "acc,none": 0.7574850299401198, + "acc_stderr,none": 0.023487359027875285 + }, + "arabicmmlu_high_philosophy": { + "alias": " - High Philosophy", + "acc,none": 0.7435897435897436, + "acc_stderr,none": 0.07083413480167725 + }, + "arabicmmlu_islamic_studies": { + "alias": " - Islamic Studies", + "acc,none": 0.7089201877934272, + "acc_stderr,none": 0.017984334664115503 + }, + "arabicmmlu_middle_history": { + "alias": " - Middle History", + "acc,none": 0.7586206896551724, + "acc_stderr,none": 0.03010833071801162 + }, + "arabicmmlu_middle_islamic_studies": { + "alias": " - Middle Islamic Studies", + "acc,none": 0.7899159663865546, + "acc_stderr,none": 0.026461398717471874 + }, + "arabicmmlu_primary_history": { + "alias": " - Primary History", + "acc,none": 0.7058823529411765, + "acc_stderr,none": 0.04533838195929775 + }, + "arabicmmlu_primary_islamic_studies": { + "alias": " - Primary Islamic Studies", + "acc,none": 0.8548548548548549, + "acc_stderr,none": 0.011150187682575276 + }, + "arabicmmlu_prof_law": { + "alias": " - Prof Law", + "acc,none": 0.767515923566879, + "acc_stderr,none": 0.023876360884096247 + }, + "arabicmmlu_language": { + "acc,none": 0.704131227217497, + "acc_stderr,none": 0.01074858647087823, + "alias": " - Language" + }, + "arabicmmlu_arabic_language_(general)": { + "alias": " - Arabic Language (General)", + "acc,none": 0.8169934640522876, + "acc_stderr,none": 0.015643069911273347 + }, + "arabicmmlu_arabic_language_(grammar)": { + "alias": " - Arabic Language (Grammar)", + "acc,none": 0.6986301369863014, + "acc_stderr,none": 0.024050431713518203 + }, + "arabicmmlu_high_arabic_language": { + "alias": " - High Arabic Language", + "acc,none": 0.4717948717948718, + "acc_stderr,none": 0.025310639254933903 + }, + "arabicmmlu_middle_arabic_language": { + "alias": " - Middle Arabic Language", + "acc,none": 0.7777777777777778, + "acc_stderr,none": 0.08153326507837146 + }, + "arabicmmlu_primary_arabic_language": { + "alias": " - Primary Arabic Language", + "acc,none": 0.7896825396825397, + "acc_stderr,none": 0.025723323024496765 + }, + "arabicmmlu_other": { + "acc,none": 0.7564412238325282, + "acc_stderr,none": 0.008605534818784389, + "alias": " - Other" + }, + "arabicmmlu_driving_test": { + "alias": " - Driving Test", + "acc,none": 0.7704376548307185, + "acc_stderr,none": 0.012090002524101525 + }, + "arabicmmlu_general_knowledge": { + "alias": " - General Knowledge", + "acc,none": 0.7245370370370371, + "acc_stderr,none": 0.015207453766372243 + }, + "arabicmmlu_middle_general_knowledge": { + "alias": " - Middle General Knowledge", + "acc,none": 0.7848837209302325, + "acc_stderr,none": 0.0314225368473594 + }, + "arabicmmlu_primary_general_knowledge": { + "alias": " - Primary General Knowledge", + "acc,none": 0.7592592592592593, + "acc_stderr,none": 0.033694336336687475 + }, + "arabicmmlu_univ_management": { + "alias": " - Univ Management", + "acc,none": 0.8266666666666667, + "acc_stderr,none": 0.04400382183783964 + }, + "arabicmmlu_social_science": { + "acc,none": 0.697203196347032, + "acc_stderr,none": 0.007663541005039597, + "alias": " - Social Science" + }, + "arabicmmlu_high_civics": { + "alias": " - High Civics", + "acc,none": 0.5977011494252874, + "acc_stderr,none": 0.052877049732218045 + }, + "arabicmmlu_high_economics": { + "alias": " - High Economics", + "acc,none": 0.7166666666666667, + "acc_stderr,none": 0.023782648315084427 + }, + "arabicmmlu_high_geography": { + "alias": " - High Geography", + "acc,none": 0.6290944123314065, + "acc_stderr,none": 0.015000309630517242 + }, + "arabicmmlu_middle_civics": { + "alias": " - Middle Civics", + "acc,none": 0.6228813559322034, + "acc_stderr,none": 0.03161605923498462 + }, + "arabicmmlu_middle_economics": { + "alias": " - Middle Economics", + "acc,none": 0.7931034482758621, + "acc_stderr,none": 0.04368097459950702 + }, + "arabicmmlu_middle_geography": { + "alias": " - Middle Geography", + "acc,none": 0.7389705882352942, + "acc_stderr,none": 0.026679252270103114 + }, + "arabicmmlu_middle_social_science": { + "alias": " - Middle Social Science", + "acc,none": 0.6390041493775933, + "acc_stderr,none": 0.031002543340279055 + }, + "arabicmmlu_primary_geography": { + "alias": " - Primary Geography", + "acc,none": 0.7368421052631579, + "acc_stderr,none": 0.058843894144731304 + }, + "arabicmmlu_primary_social_science": { + "alias": " - Primary Social Science", + "acc,none": 0.825531914893617, + "acc_stderr,none": 0.014303377520795746 + }, + "arabicmmlu_univ_accounting": { + "alias": " - Univ Accounting", + "acc,none": 0.6621621621621622, + "acc_stderr,none": 0.05535729934952123 + }, + "arabicmmlu_univ_economics": { + "alias": " - Univ Economics", + "acc,none": 0.6715328467153284, + "acc_stderr,none": 0.04027264457070886 + }, + "arabicmmlu_univ_political_science": { + "alias": " - Univ Political Science", + "acc,none": 0.6857142857142857, + "acc_stderr,none": 0.0321115135399438 + }, + "arabicmmlu_stem": { + "acc,none": 0.7062323833385531, + "acc_stderr,none": 0.007870570600880707, + "alias": " - STEM" + }, + "arabicmmlu_high_biology": { + "alias": " - High Biology", + "acc,none": 0.6153300212916962, + "acc_stderr,none": 0.012965726952941084 + }, + "arabicmmlu_high_computer_science": { + "alias": " - High Computer Science", + "acc,none": 0.7471264367816092, + "acc_stderr,none": 0.026956412412778324 + }, + "arabicmmlu_high_physics": { + "alias": " - High Physics", + "acc,none": 0.6509803921568628, + "acc_stderr,none": 0.029908319306125593 + }, + "arabicmmlu_middle_computer_science": { + "alias": " - Middle Computer Science", + "acc,none": 0.9629629629629629, + "acc_stderr,none": 0.03703703703703703 + }, + "arabicmmlu_middle_natural_science": { + "alias": " - Middle Natural Science", + "acc,none": 0.8429752066115702, + "acc_stderr,none": 0.023435973310697193 + }, + "arabicmmlu_primary_computer_science": { + "alias": " - Primary Computer Science", + "acc,none": 0.7789473684210526, + "acc_stderr,none": 0.030183597428219758 + }, + "arabicmmlu_primary_math": { + "alias": " - Primary Math", + "acc,none": 0.7334963325183375, + "acc_stderr,none": 0.02188872609697175 + }, + "arabicmmlu_primary_natural_science": { + "alias": " - Primary Natural Science", + "acc,none": 0.8958333333333334, + "acc_stderr,none": 0.016689971269054218 + }, + "arabicmmlu_univ_computer_science": { + "alias": " - Univ Computer Science", + "acc,none": 0.75, + "acc_stderr,none": 0.05455447255899809 + } + }, + "groups": { + "arabicmmlu": { + "acc,none": 0.7200968523002421, + "acc_stderr,none": 0.003653809830387355, + "alias": "arabicmmlu" + }, + "arabicmmlu_humanities": { + "acc,none": 0.7367695700110254, + "acc_stderr,none": 0.007118478408616655, + "alias": " - Humanities" + }, + "arabicmmlu_language": { + "acc,none": 0.704131227217497, + "acc_stderr,none": 0.01074858647087823, + "alias": " - Language" + }, + "arabicmmlu_other": { + "acc,none": 0.7564412238325282, + "acc_stderr,none": 0.008605534818784389, + "alias": " - Other" + }, + "arabicmmlu_social_science": { + "acc,none": 0.697203196347032, + "acc_stderr,none": 0.007663541005039597, + "alias": " - Social Science" + }, + "arabicmmlu_stem": { + "acc,none": 0.7062323833385531, + "acc_stderr,none": 0.007870570600880707, + "alias": " - STEM" + } + }, + "group_subtasks": { + "arabicmmlu_language": [ + "arabicmmlu_arabic_language_(general)", + "arabicmmlu_primary_arabic_language", + "arabicmmlu_middle_arabic_language", + "arabicmmlu_high_arabic_language", + "arabicmmlu_arabic_language_(grammar)" + ], + "arabicmmlu_stem": [ + "arabicmmlu_middle_computer_science", + "arabicmmlu_primary_math", + "arabicmmlu_primary_natural_science", + "arabicmmlu_high_biology", + "arabicmmlu_middle_natural_science", + "arabicmmlu_high_physics", + "arabicmmlu_high_computer_science", + "arabicmmlu_univ_computer_science", + "arabicmmlu_primary_computer_science" + ], + "arabicmmlu_humanities": [ + "arabicmmlu_high_history", + "arabicmmlu_middle_history", + "arabicmmlu_high_philosophy", + "arabicmmlu_prof_law", + "arabicmmlu_primary_islamic_studies", + "arabicmmlu_high_islamic_studies", + "arabicmmlu_primary_history", + "arabicmmlu_middle_islamic_studies", + "arabicmmlu_islamic_studies" + ], + "arabicmmlu_social_science": [ + "arabicmmlu_high_civics", + "arabicmmlu_univ_political_science", + "arabicmmlu_high_economics", + "arabicmmlu_middle_economics", + "arabicmmlu_univ_economics", + "arabicmmlu_high_geography", + "arabicmmlu_primary_geography", + "arabicmmlu_middle_civics", + "arabicmmlu_univ_accounting", + "arabicmmlu_middle_social_science", + "arabicmmlu_middle_geography", + "arabicmmlu_primary_social_science" + ], + "arabicmmlu_other": [ + "arabicmmlu_general_knowledge", + "arabicmmlu_middle_general_knowledge", + "arabicmmlu_primary_general_knowledge", + "arabicmmlu_univ_management", + "arabicmmlu_driving_test" + ], + "arabicmmlu": [ + "arabicmmlu_other", + "arabicmmlu_social_science", + "arabicmmlu_humanities", + "arabicmmlu_stem", + "arabicmmlu_language" + ] + }, + "configs": { + "arabicmmlu_arabic_language_(general)": { + "task": "arabicmmlu_arabic_language_(general)", + "task_alias": "Arabic Language (General)", + "tag": "arabicmmlu_language_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Arabic Language (General)", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_arabic_language_(grammar)": { + "task": "arabicmmlu_arabic_language_(grammar)", + "task_alias": "Arabic Language (Grammar)", + "tag": "arabicmmlu_language_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Arabic Language (Grammar)", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_driving_test": { + "task": "arabicmmlu_driving_test", + "task_alias": "Driving Test", + "tag": "arabicmmlu_other_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Driving Test", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_general_knowledge": { + "task": "arabicmmlu_general_knowledge", + "task_alias": "General Knowledge", + "tag": "arabicmmlu_other_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "General Knowledge", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_arabic_language": { + "task": "arabicmmlu_high_arabic_language", + "task_alias": "High Arabic Language", + "tag": "arabicmmlu_language_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Arabic Language", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_biology": { + "task": "arabicmmlu_high_biology", + "task_alias": "High Biology", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Biology", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_civics": { + "task": "arabicmmlu_high_civics", + "task_alias": "High Civics", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Civics", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_computer_science": { + "task": "arabicmmlu_high_computer_science", + "task_alias": "High Computer Science", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Computer Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_economics": { + "task": "arabicmmlu_high_economics", + "task_alias": "High Economics", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Economics", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_geography": { + "task": "arabicmmlu_high_geography", + "task_alias": "High Geography", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Geography", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_history": { + "task": "arabicmmlu_high_history", + "task_alias": "High History", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High History", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_islamic_studies": { + "task": "arabicmmlu_high_islamic_studies", + "task_alias": "High Islamic Studies", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Islamic Studies", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_philosophy": { + "task": "arabicmmlu_high_philosophy", + "task_alias": "High Philosophy", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Philosophy", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_high_physics": { + "task": "arabicmmlu_high_physics", + "task_alias": "High Physics", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "High Physics", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_islamic_studies": { + "task": "arabicmmlu_islamic_studies", + "task_alias": "Islamic Studies", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Islamic Studies", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_arabic_language": { + "task": "arabicmmlu_middle_arabic_language", + "task_alias": "Middle Arabic Language", + "tag": "arabicmmlu_language_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Arabic Language", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_civics": { + "task": "arabicmmlu_middle_civics", + "task_alias": "Middle Civics", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Civics", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_computer_science": { + "task": "arabicmmlu_middle_computer_science", + "task_alias": "Middle Computer Science", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Computer Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_economics": { + "task": "arabicmmlu_middle_economics", + "task_alias": "Middle Economics", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Economics", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_general_knowledge": { + "task": "arabicmmlu_middle_general_knowledge", + "task_alias": "Middle General Knowledge", + "tag": "arabicmmlu_other_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle General Knowledge", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_geography": { + "task": "arabicmmlu_middle_geography", + "task_alias": "Middle Geography", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Geography", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_history": { + "task": "arabicmmlu_middle_history", + "task_alias": "Middle History", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle History", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_islamic_studies": { + "task": "arabicmmlu_middle_islamic_studies", + "task_alias": "Middle Islamic Studies", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Islamic Studies", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_natural_science": { + "task": "arabicmmlu_middle_natural_science", + "task_alias": "Middle Natural Science", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Natural Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_middle_social_science": { + "task": "arabicmmlu_middle_social_science", + "task_alias": "Middle Social Science", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Middle Social Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_arabic_language": { + "task": "arabicmmlu_primary_arabic_language", + "task_alias": "Primary Arabic Language", + "tag": "arabicmmlu_language_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Arabic Language", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_computer_science": { + "task": "arabicmmlu_primary_computer_science", + "task_alias": "Primary Computer Science", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Computer Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_general_knowledge": { + "task": "arabicmmlu_primary_general_knowledge", + "task_alias": "Primary General Knowledge", + "tag": "arabicmmlu_other_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary General Knowledge", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_geography": { + "task": "arabicmmlu_primary_geography", + "task_alias": "Primary Geography", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Geography", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_history": { + "task": "arabicmmlu_primary_history", + "task_alias": "Primary History", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary History", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_islamic_studies": { + "task": "arabicmmlu_primary_islamic_studies", + "task_alias": "Primary Islamic Studies", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Islamic Studies", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_math": { + "task": "arabicmmlu_primary_math", + "task_alias": "Primary Math", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Math", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_natural_science": { + "task": "arabicmmlu_primary_natural_science", + "task_alias": "Primary Natural Science", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Natural Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_primary_social_science": { + "task": "arabicmmlu_primary_social_science", + "task_alias": "Primary Social Science", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Primary Social Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_prof_law": { + "task": "arabicmmlu_prof_law", + "task_alias": "Prof Law", + "tag": "arabicmmlu_humanities_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Prof Law", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_univ_accounting": { + "task": "arabicmmlu_univ_accounting", + "task_alias": "Univ Accounting", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Univ Accounting", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_univ_computer_science": { + "task": "arabicmmlu_univ_computer_science", + "task_alias": "Univ Computer Science", + "tag": "arabicmmlu_stem_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Univ Computer Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_univ_economics": { + "task": "arabicmmlu_univ_economics", + "task_alias": "Univ Economics", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Univ Economics", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_univ_management": { + "task": "arabicmmlu_univ_management", + "task_alias": "Univ Management", + "tag": "arabicmmlu_other_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Univ Management", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + }, + "arabicmmlu_univ_political_science": { + "task": "arabicmmlu_univ_political_science", + "task_alias": "Univ Political Science", + "tag": "arabicmmlu_social_science_tasks", + "dataset_path": "yazeed7/ArabicMMLU", + "dataset_name": "Univ Political Science", + "test_split": "test", + "fewshot_split": "dev", + "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", + "doc_to_target": "Answer Key", + "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", + "description": "", + "target_delimiter": " ", + "fewshot_delimiter": "\n\n", + "fewshot_config": { + "sampler": "first_n" + }, + "num_fewshot": 0, + "metric_list": [ + { + "metric": "acc", + "aggregation": "mean", + "higher_is_better": true + } + ], + "output_type": "multiple_choice", + "repeats": 1, + "should_decontaminate": false, + "metadata": { + "version": 0.0 + } + } + }, + "versions": { + "arabicmmlu": 0, + "arabicmmlu_arabic_language_(general)": 0.0, + "arabicmmlu_arabic_language_(grammar)": 0.0, + "arabicmmlu_driving_test": 0.0, + "arabicmmlu_general_knowledge": 0.0, + "arabicmmlu_high_arabic_language": 0.0, + "arabicmmlu_high_biology": 0.0, + "arabicmmlu_high_civics": 0.0, + "arabicmmlu_high_computer_science": 0.0, + "arabicmmlu_high_economics": 0.0, + "arabicmmlu_high_geography": 0.0, + "arabicmmlu_high_history": 0.0, + "arabicmmlu_high_islamic_studies": 0.0, + "arabicmmlu_high_philosophy": 0.0, + "arabicmmlu_high_physics": 0.0, + "arabicmmlu_humanities": 0, + "arabicmmlu_islamic_studies": 0.0, + "arabicmmlu_language": 0, + "arabicmmlu_middle_arabic_language": 0.0, + "arabicmmlu_middle_civics": 0.0, + "arabicmmlu_middle_computer_science": 0.0, + "arabicmmlu_middle_economics": 0.0, + "arabicmmlu_middle_general_knowledge": 0.0, + "arabicmmlu_middle_geography": 0.0, + "arabicmmlu_middle_history": 0.0, + "arabicmmlu_middle_islamic_studies": 0.0, + "arabicmmlu_middle_natural_science": 0.0, + "arabicmmlu_middle_social_science": 0.0, + "arabicmmlu_other": 0, + "arabicmmlu_primary_arabic_language": 0.0, + "arabicmmlu_primary_computer_science": 0.0, + "arabicmmlu_primary_general_knowledge": 0.0, + "arabicmmlu_primary_geography": 0.0, + "arabicmmlu_primary_history": 0.0, + "arabicmmlu_primary_islamic_studies": 0.0, + "arabicmmlu_primary_math": 0.0, + "arabicmmlu_primary_natural_science": 0.0, + "arabicmmlu_primary_social_science": 0.0, + "arabicmmlu_prof_law": 0.0, + "arabicmmlu_social_science": 0, + "arabicmmlu_stem": 0, + "arabicmmlu_univ_accounting": 0.0, + "arabicmmlu_univ_computer_science": 0.0, + "arabicmmlu_univ_economics": 0.0, + "arabicmmlu_univ_management": 0.0, + "arabicmmlu_univ_political_science": 0.0 + }, + "n-shot": { + "arabicmmlu_arabic_language_(general)": 0, + "arabicmmlu_arabic_language_(grammar)": 0, + "arabicmmlu_driving_test": 0, + "arabicmmlu_general_knowledge": 0, + "arabicmmlu_high_arabic_language": 0, + "arabicmmlu_high_biology": 0, + "arabicmmlu_high_civics": 0, + "arabicmmlu_high_computer_science": 0, + "arabicmmlu_high_economics": 0, + "arabicmmlu_high_geography": 0, + "arabicmmlu_high_history": 0, + "arabicmmlu_high_islamic_studies": 0, + "arabicmmlu_high_philosophy": 0, + "arabicmmlu_high_physics": 0, + "arabicmmlu_islamic_studies": 0, + "arabicmmlu_middle_arabic_language": 0, + "arabicmmlu_middle_civics": 0, + "arabicmmlu_middle_computer_science": 0, + "arabicmmlu_middle_economics": 0, + "arabicmmlu_middle_general_knowledge": 0, + "arabicmmlu_middle_geography": 0, + "arabicmmlu_middle_history": 0, + "arabicmmlu_middle_islamic_studies": 0, + "arabicmmlu_middle_natural_science": 0, + "arabicmmlu_middle_social_science": 0, + "arabicmmlu_primary_arabic_language": 0, + "arabicmmlu_primary_computer_science": 0, + "arabicmmlu_primary_general_knowledge": 0, + "arabicmmlu_primary_geography": 0, + "arabicmmlu_primary_history": 0, + "arabicmmlu_primary_islamic_studies": 0, + "arabicmmlu_primary_math": 0, + "arabicmmlu_primary_natural_science": 0, + "arabicmmlu_primary_social_science": 0, + "arabicmmlu_prof_law": 0, + "arabicmmlu_univ_accounting": 0, + "arabicmmlu_univ_computer_science": 0, + "arabicmmlu_univ_economics": 0, + "arabicmmlu_univ_management": 0, + "arabicmmlu_univ_political_science": 0 + }, + "higher_is_better": { + "arabicmmlu": { + "acc": true + }, + "arabicmmlu_arabic_language_(general)": { + "acc": true + }, + "arabicmmlu_arabic_language_(grammar)": { + "acc": true + }, + "arabicmmlu_driving_test": { + "acc": true + }, + "arabicmmlu_general_knowledge": { + "acc": true + }, + "arabicmmlu_high_arabic_language": { + "acc": true + }, + "arabicmmlu_high_biology": { + "acc": true + }, + "arabicmmlu_high_civics": { + "acc": true + }, + "arabicmmlu_high_computer_science": { + "acc": true + }, + "arabicmmlu_high_economics": { + "acc": true + }, + "arabicmmlu_high_geography": { + "acc": true + }, + "arabicmmlu_high_history": { + "acc": true + }, + "arabicmmlu_high_islamic_studies": { + "acc": true + }, + "arabicmmlu_high_philosophy": { + "acc": true + }, + "arabicmmlu_high_physics": { + "acc": true + }, + "arabicmmlu_humanities": { + "acc": true + }, + "arabicmmlu_islamic_studies": { + "acc": true + }, + "arabicmmlu_language": { + "acc": true + }, + "arabicmmlu_middle_arabic_language": { + "acc": true + }, + "arabicmmlu_middle_civics": { + "acc": true + }, + "arabicmmlu_middle_computer_science": { + "acc": true + }, + "arabicmmlu_middle_economics": { + "acc": true + }, + "arabicmmlu_middle_general_knowledge": { + "acc": true + }, + "arabicmmlu_middle_geography": { + "acc": true + }, + "arabicmmlu_middle_history": { + "acc": true + }, + "arabicmmlu_middle_islamic_studies": { + "acc": true + }, + "arabicmmlu_middle_natural_science": { + "acc": true + }, + "arabicmmlu_middle_social_science": { + "acc": true + }, + "arabicmmlu_other": { + "acc": true + }, + "arabicmmlu_primary_arabic_language": { + "acc": true + }, + "arabicmmlu_primary_computer_science": { + "acc": true + }, + "arabicmmlu_primary_general_knowledge": { + "acc": true + }, + "arabicmmlu_primary_geography": { + "acc": true + }, + "arabicmmlu_primary_history": { + "acc": true + }, + "arabicmmlu_primary_islamic_studies": { + "acc": true + }, + "arabicmmlu_primary_math": { + "acc": true + }, + "arabicmmlu_primary_natural_science": { + "acc": true + }, + "arabicmmlu_primary_social_science": { + "acc": true + }, + "arabicmmlu_prof_law": { + "acc": true + }, + "arabicmmlu_social_science": { + "acc": true + }, + "arabicmmlu_stem": { + "acc": true + }, + "arabicmmlu_univ_accounting": { + "acc": true + }, + "arabicmmlu_univ_computer_science": { + "acc": true + }, + "arabicmmlu_univ_economics": { + "acc": true + }, + "arabicmmlu_univ_management": { + "acc": true + }, + "arabicmmlu_univ_political_science": { + "acc": true + } + }, + "n-samples": { + "arabicmmlu_general_knowledge": { + "original": 864, + "effective": 864 + }, + "arabicmmlu_middle_general_knowledge": { + "original": 172, + "effective": 172 + }, + "arabicmmlu_primary_general_knowledge": { + "original": 162, + "effective": 162 + }, + "arabicmmlu_univ_management": { + "original": 75, + "effective": 75 + }, + "arabicmmlu_driving_test": { + "original": 1211, + "effective": 1211 + }, + "arabicmmlu_high_civics": { + "original": 87, + "effective": 87 + }, + "arabicmmlu_univ_political_science": { + "original": 210, + "effective": 210 + }, + "arabicmmlu_high_economics": { + "original": 360, + "effective": 360 + }, + "arabicmmlu_middle_economics": { + "original": 87, + "effective": 87 + }, + "arabicmmlu_univ_economics": { + "original": 137, + "effective": 137 + }, + "arabicmmlu_high_geography": { + "original": 1038, + "effective": 1038 + }, + "arabicmmlu_primary_geography": { + "original": 57, + "effective": 57 + }, + "arabicmmlu_middle_civics": { + "original": 236, + "effective": 236 + }, + "arabicmmlu_univ_accounting": { + "original": 74, + "effective": 74 + }, + "arabicmmlu_middle_social_science": { + "original": 241, + "effective": 241 + }, + "arabicmmlu_middle_geography": { + "original": 272, + "effective": 272 + }, + "arabicmmlu_primary_social_science": { + "original": 705, + "effective": 705 + }, + "arabicmmlu_high_history": { + "original": 760, + "effective": 760 + }, + "arabicmmlu_middle_history": { + "original": 203, + "effective": 203 + }, + "arabicmmlu_high_philosophy": { + "original": 39, + "effective": 39 + }, + "arabicmmlu_prof_law": { + "original": 314, + "effective": 314 + }, + "arabicmmlu_primary_islamic_studies": { + "original": 999, + "effective": 999 + }, + "arabicmmlu_high_islamic_studies": { + "original": 334, + "effective": 334 + }, + "arabicmmlu_primary_history": { + "original": 102, + "effective": 102 + }, + "arabicmmlu_middle_islamic_studies": { + "original": 238, + "effective": 238 + }, + "arabicmmlu_islamic_studies": { + "original": 639, + "effective": 639 + }, + "arabicmmlu_middle_computer_science": { + "original": 27, + "effective": 27 + }, + "arabicmmlu_primary_math": { + "original": 409, + "effective": 409 + }, + "arabicmmlu_primary_natural_science": { + "original": 336, + "effective": 336 + }, + "arabicmmlu_high_biology": { + "original": 1409, + "effective": 1409 + }, + "arabicmmlu_middle_natural_science": { + "original": 242, + "effective": 242 + }, + "arabicmmlu_high_physics": { + "original": 255, + "effective": 255 + }, + "arabicmmlu_high_computer_science": { + "original": 261, + "effective": 261 + }, + "arabicmmlu_univ_computer_science": { + "original": 64, + "effective": 64 + }, + "arabicmmlu_primary_computer_science": { + "original": 190, + "effective": 190 + }, + "arabicmmlu_arabic_language_(general)": { + "original": 612, + "effective": 612 + }, + "arabicmmlu_primary_arabic_language": { + "original": 252, + "effective": 252 + }, + "arabicmmlu_middle_arabic_language": { + "original": 27, + "effective": 27 + }, + "arabicmmlu_high_arabic_language": { + "original": 390, + "effective": 390 + }, + "arabicmmlu_arabic_language_(grammar)": { + "original": 365, + "effective": 365 + } + }, + "config": { + "model": "hf", + "model_args": "pretrained=meta-llama/Llama-3.3-70B-Instruct,trust_remote_code=True,cache_dir=/tmp,parallelize=True", + "model_num_parameters": 70553706496, + "model_dtype": "torch.bfloat16", + "model_revision": "main", + "model_sha": "6f6073b423013f6a7d4d9f39144961bfbfbc386b", + "batch_size": "auto", + "batch_sizes": [ + 16 + ], + "device": null, + "use_cache": null, + "limit": null, + "bootstrap_iters": 100000, + "gen_kwargs": null, + "random_seed": 0, + "numpy_seed": 1234, + "torch_seed": 1234, + "fewshot_seed": 1234 + }, + "git_hash": "788a3672", + "date": 1737858946.4669714, + "pretty_env_info": "PyTorch version: 2.4.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: Ubuntu 22.04.3 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0\nClang version: Could not collect\nCMake version: version 3.27.1\nLibc version: glibc-2.35\n\nPython version: 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] (64-bit runtime)\nPython platform: Linux-5.15.0-1064-azure-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.2.128\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: \nGPU 0: NVIDIA A100-SXM4-80GB\nGPU 1: NVIDIA A100-SXM4-80GB\nGPU 2: NVIDIA A100-SXM4-80GB\nGPU 3: NVIDIA A100-SXM4-80GB\nGPU 4: NVIDIA A100-SXM4-80GB\nGPU 5: NVIDIA A100-SXM4-80GB\nGPU 6: NVIDIA A100-SXM4-80GB\nGPU 7: NVIDIA A100-SXM4-80GB\n\nNvidia driver version: 535.161.08\ncuDNN version: Probably one of the following:\n/usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.4\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nAddress sizes: 48 bits physical, 48 bits virtual\nByte Order: Little Endian\nCPU(s): 96\nOn-line CPU(s) list: 0-95\nVendor ID: AuthenticAMD\nModel name: AMD EPYC 7V12 64-Core Processor\nCPU family: 23\nModel: 49\nThread(s) per core: 1\nCore(s) per socket: 48\nSocket(s): 2\nStepping: 0\nBogoMIPS: 4890.89\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 tsc_reliable nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw topoext perfctr_core ssbd vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr rdpru arat umip rdpid\nHypervisor vendor: Microsoft\nVirtualization type: full\nL1d cache: 3 MiB (96 instances)\nL1i cache: 3 MiB (96 instances)\nL2 cache: 48 MiB (96 instances)\nL3 cache: 384 MiB (24 instances)\nNUMA node(s): 4\nNUMA node0 CPU(s): 0-23\nNUMA node1 CPU(s): 24-47\nNUMA node2 CPU(s): 48-71\nNUMA node3 CPU(s): 72-95\nVulnerability Gather data sampling: 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 Retbleed: Mitigation; untrained return thunk; SMT disabled\nVulnerability Spec rstack overflow: Mitigation; safe RET, no microcode\nVulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines; STIBP disabled; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] onnx==1.14.0\n[pip3] pytorch-lightning==2.0.7\n[pip3] pytorch-quantization==2.1.2\n[pip3] torch==2.4.0\n[pip3] torch-tensorrt==2.0.0.dev0\n[pip3] torchaudio==2.1.0\n[pip3] torchdata==0.7.0a0\n[pip3] torchmetrics==1.2.0\n[pip3] torchvision==0.19.0\n[pip3] triton==3.0.0\n[conda] Could not collect", + "transformers_version": "4.48.1", + "upper_git_hash": "086919bd66f4e15fdcd4b792a7b27a698c1ba091", + "tokenizer_pad_token": [ + "<|finetune_right_pad_id|>", + "128004" + ], + "tokenizer_eos_token": [ + "<|eot_id|>", + "128009" + ], + "tokenizer_bos_token": [ + "<|begin_of_text|>", + "128000" + ], + "eot_token_id": 128009, + "max_length": 131072, + "task_hashes": {}, + "model_source": "hf", + "model_name": "meta-llama/Llama-3.3-70B-Instruct", + "model_name_sanitized": "meta-llama__Llama-3.3-70B-Instruct", + "system_instruction": null, + "system_instruction_sha": null, + "fewshot_as_multiturn": false, + "chat_template": null, + "chat_template_sha": null, + "start_time": 820233.226282937, + "end_time": 821135.688521802, + "total_evaluation_time_seconds": "902.4622388649732" +} \ No newline at end of file