five

open-llm-leaderboard/details_facebook__opt-iml-max-1.3b

收藏
Hugging Face2023-10-18 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_facebook__opt-iml-max-1.3b
下载链接
链接失效反馈
官方服务:
资源简介:
--- pretty_name: Evaluation run of facebook/opt-iml-max-1.3b dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [facebook/opt-iml-max-1.3b](https://huggingface.co/facebook/opt-iml-max-1.3b)\ \ on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).\n\ \nThe dataset is composed of 64 configuration, each one coresponding to one of the\ \ evaluated task.\n\nThe dataset has been created from 2 run(s). Each run can be\ \ found as a specific split in each configuration, the split being named using the\ \ timestamp of the run.The \"train\" split is always pointing to the latest results.\n\ \nAn additional configuration \"results\" store all the aggregated results of the\ \ run (and is used to compute and display the agregated metrics on the [Open LLM\ \ Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)).\n\ \nTo load the details from a run, you can for instance do the following:\n```python\n\ from datasets import load_dataset\ndata = load_dataset(\"open-llm-leaderboard/details_facebook__opt-iml-max-1.3b\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-10-18T09:50:43.719660](https://huggingface.co/datasets/open-llm-leaderboard/details_facebook__opt-iml-max-1.3b/blob/main/results_2023-10-18T09-50-43.719660.json)(note\ \ that their might be results for other tasks in the repos if successive evals didn't\ \ cover the same tasks. You find each in the results and the \"latest\" split for\ \ each eval):\n\n```python\n{\n \"all\": {\n \"em\": 0.3028523489932886,\n\ \ \"em_stderr\": 0.0047056271048806315,\n \"f1\": 0.3369934983221478,\n\ \ \"f1_stderr\": 0.004663613383395755,\n \"acc\": 0.30375849777371944,\n\ \ \"acc_stderr\": 0.007878524617348554\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.3028523489932886,\n \"em_stderr\": 0.0047056271048806315,\n\ \ \"f1\": 0.3369934983221478,\n \"f1_stderr\": 0.004663613383395755\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.00530705079605762,\n \ \ \"acc_stderr\": 0.0020013057209480856\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.6022099447513812,\n \"acc_stderr\": 0.013755743513749023\n\ \ }\n}\n```" repo_url: https://huggingface.co/facebook/opt-iml-max-1.3b leaderboard_url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard point_of_contact: clementine@hf.co configs: - config_name: harness_arc_challenge_25 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|arc:challenge|25_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-08-22T09:51:53.668877.parquet' - config_name: harness_drop_3 data_files: - split: 2023_10_18T09_50_43.719660 path: - '**/details_harness|drop|3_2023-10-18T09-50-43.719660.parquet' - split: latest path: - '**/details_harness|drop|3_2023-10-18T09-50-43.719660.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_10_18T09_50_43.719660 path: - '**/details_harness|gsm8k|5_2023-10-18T09-50-43.719660.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-10-18T09-50-43.719660.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hellaswag|10_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-management|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-management|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-08-22T09:51:53.668877.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-international_law|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-management|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-marketing|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-sociology|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-virology|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-08-22T09:51:53.668877.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_08_22T09_51_53.668877 path: - '**/details_harness|truthfulqa:mc|0_2023-08-22T09:51:53.668877.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-08-22T09:51:53.668877.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_10_18T09_50_43.719660 path: - '**/details_harness|winogrande|5_2023-10-18T09-50-43.719660.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-10-18T09-50-43.719660.parquet' - config_name: results data_files: - split: 2023_10_18T09_50_43.719660 path: - results_2023-10-18T09-50-43.719660.parquet - split: latest path: - results_2023-10-18T09-50-43.719660.parquet --- # Dataset Card for Evaluation run of facebook/opt-iml-max-1.3b ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/facebook/opt-iml-max-1.3b - **Paper:** - **Leaderboard:** https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard - **Point of Contact:** clementine@hf.co ### Dataset Summary Dataset automatically created during the evaluation run of model [facebook/opt-iml-max-1.3b](https://huggingface.co/facebook/opt-iml-max-1.3b) on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). The dataset is composed of 64 configuration, each one coresponding to one of the evaluated task. The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results. An additional configuration "results" store all the aggregated results of the run (and is used to compute and display the agregated metrics on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)). To load the details from a run, you can for instance do the following: ```python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_facebook__opt-iml-max-1.3b", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-10-18T09:50:43.719660](https://huggingface.co/datasets/open-llm-leaderboard/details_facebook__opt-iml-max-1.3b/blob/main/results_2023-10-18T09-50-43.719660.json)(note that their might be results for other tasks in the repos if successive evals didn't cover the same tasks. You find each in the results and the "latest" split for each eval): ```python { "all": { "em": 0.3028523489932886, "em_stderr": 0.0047056271048806315, "f1": 0.3369934983221478, "f1_stderr": 0.004663613383395755, "acc": 0.30375849777371944, "acc_stderr": 0.007878524617348554 }, "harness|drop|3": { "em": 0.3028523489932886, "em_stderr": 0.0047056271048806315, "f1": 0.3369934983221478, "f1_stderr": 0.004663613383395755 }, "harness|gsm8k|5": { "acc": 0.00530705079605762, "acc_stderr": 0.0020013057209480856 }, "harness|winogrande|5": { "acc": 0.6022099447513812, "acc_stderr": 0.013755743513749023 } } ``` ### Supported Tasks and Leaderboards [More Information Needed] ### Languages [More Information Needed] ## Dataset Structure ### Data Instances [More Information Needed] ### Data Fields [More Information Needed] ### Data Splits [More Information Needed] ## Dataset Creation ### Curation Rationale [More Information Needed] ### Source Data #### Initial Data Collection and Normalization [More Information Needed] #### Who are the source language producers? [More Information Needed] ### Annotations #### Annotation process [More Information Needed] #### Who are the annotators? [More Information Needed] ### Personal and Sensitive Information [More Information Needed] ## Considerations for Using the Data ### Social Impact of Dataset [More Information Needed] ### Discussion of Biases [More Information Needed] ### Other Known Limitations [More Information Needed] ## Additional Information ### Dataset Curators [More Information Needed] ### Licensing Information [More Information Needed] ### Citation Information [More Information Needed] ### Contributions [More Information Needed]
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集简介

该数据集是在评估模型 facebook/opt-iml-max-1.3bOpen LLM Leaderboard 上的运行过程中自动创建的。

数据集结构

  • 配置数量:64个配置,每个配置对应一个评估任务。
  • 创建来源:数据集从2次运行中创建。每次运行可以在每个配置中找到特定的分割,分割名称使用运行的时间戳。
  • 最新结果:"train" 分割始终指向最新的结果。
  • 结果汇总:一个额外的配置 "results" 存储所有运行的汇总结果,用于计算和显示在 Open LLM Leaderboard 上的聚合指标。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_facebook__opt-iml-max-1.3b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-18T09:50:43.719660 运行的最新结果: python { "all": { "em": 0.3028523489932886, "em_stderr": 0.0047056271048806315, "f1": 0.3369934983221478, "f1_stderr": 0.004663613383395755, "acc": 0.30375849777371944, "acc_stderr": 0.007878524617348554 }, "harness|drop|3": { "em": 0.3028523489932886, "em_stderr": 0.0047056271048806315, "f1": 0.3369934983221478, "f1_stderr": 0.004663613383395755 }, "harness|gsm8k|5": { "acc": 0.00530705079605762, "acc_stderr": 0.0020013057209480856 }, "harness|winogrande|5": { "acc": 0.6022099447513812, "acc_stderr": 0.013755743513749023 } }

配置详情

以下是部分配置及其数据文件路径:

  • harness_arc_challenge_25

    • 分割:2023_08_22T09_51_53.668877
      • 路径:**/details_harness|arc:challenge|25_2023-08-22T09:51:53.668877.parquet
    • 分割:latest
      • 路径:**/details_harness|arc:challenge|25_2023-08-22T09:51:53.668877.parquet
  • harness_drop_3

    • 分割:2023_10_18T09_50_43.719660
      • 路径:**/details_harness|drop|3_2023-10-18T09-50-43.719660.parquet
    • 分割:latest
      • 路径:**/details_harness|drop|3_2023-10-18T09-50-43.719660.parquet
  • harness_gsm8k_5

    • 分割:2023_10_18T09_50_43.719660
      • 路径:**/details_harness|gsm8k|5_2023-10-18T09-50-43.719660.parquet
    • 分割:latest
      • 路径:**/details_harness|gsm8k|5_2023-10-18T09-50-43.719660.parquet
  • harness_hellaswag_10

    • 分割:2023_08_22T09_51_53.668877
      • 路径:**/details_harness|hellaswag|10_2023-08-22T09:51:53.668877.parquet
    • 分割:latest
      • 路径:**/details_harness|hellaswag|10_2023-08-22T09:51:53.668877.parquet
  • harness_hendrycksTest_5

    • 分割:2023_08_22T09_51_53.668877
      • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-formal_logic|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-global_facts|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_biology|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_geography|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_physics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-human_aging|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-human_sexuality|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-international_law|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-jurisprudence|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-machine_learning|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-management|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-marketing|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-medical_genetics|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-miscellaneous|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-moral_disputes|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-nutrition|5_2023-08-22T09:51:53.668877.parquet
      • 路径:**/details_harness|hendrycksTest-philosophy|5_2023-08-22T09:51:53.668877.parquet
      • 路径:`**/details_harness|hendrycksTest-prehistory|5
搜集汇总
数据集介绍
main_image_url
构建方式
在大型语言模型的评估研究中,对模型性能的系统性评测至关重要。该数据集源自对facebook/opt-iml-max-1.3b模型在Open LLM Leaderboard上的自动化评估流程。其构建方式是通过将模型在多个标准化任务上的推理结果进行结构化存储,每个任务对应一个独立的配置(configuration),共计64个配置。每次评估运行均生成一个以时间戳命名的数据分割(split),而最新的结果则统一归入'train'分割。此外,数据集还包含一个名为'results'的配置,用于汇总所有评估运行的聚合指标。所有数据以Parquet格式存储,确保了高效的读写性能。
使用方法
研究者可以通过HuggingFace的datasets库便捷地加载该数据集。具体而言,使用load_dataset函数时需指定数据集名称和目标任务的配置名称(如'harness_winogrande_5'),并通过split参数选择所需的数据分割。例如,加载最新评估结果时可设置split='train',而若要访问特定历史运行的数据,则可使用对应的时间戳字符串作为分割标识。这种设计使得研究者能够灵活地在不同任务和不同版本之间切换,从而支持深入的模型性能剖析与复现验证工作。
背景与挑战
背景概述
随着大规模语言模型在自然语言处理领域展现出卓越的能力,如何系统性地评估其性能成为了一个关键议题。在此背景下,Hugging Face社区于2023年发起了Open LLM Leaderboard项目,旨在为研究人员提供一个标准化、透明化的模型评测平台。该数据集作为Open LLM Leaderboard的一部分,专门记录了Meta AI研发的opt-iml-max-1.3b模型的评测结果。opt-iml-max-1.3b是OPT系列中经过指令微调的变体,其核心研究问题在于探究中等规模模型在经过针对性训练后,在多种自然语言理解与推理任务上的表现边界。该数据集由Hugging Face团队负责维护,通过自动化流程收集了模型在ARC、HellaSwag、MMLU、TruthfulQA等多项基准测试上的详细结果,为社区提供了宝贵的模型性能参考,推动了开源大模型评测体系的建立与发展。
当前挑战
该数据集所应对的核心领域挑战在于,大语言模型的性能评估缺乏统一标准和可重复性,不同研究机构采用的评测环境、提示模板和指标计算方式各异,导致结果难以横向比较。构建过程中,数据集面临多重技术挑战:首先,需要设计一套兼容多种模型架构和任务的评测框架,确保评测流程的标准化;其次,面对opt-iml-max-1.3b这类中等规模模型,如何选择恰当的基准测试以全面反映其能力,避免评测偏向特定任务;最后,数据集的自动化构建流程需处理评测运行中的版本控制问题,确保每次运行结果的可追溯性,同时维护海量细粒度评测数据(涵盖64个配置和57个学科子任务)的完整性与一致性,这对数据存储和检索架构提出了较高要求。
常用场景
经典使用场景
该数据集是Open LLM Leaderboard对facebook/opt-iml-max-1.3b模型进行自动化评估过程中生成的评测记录集合,涵盖了ARC挑战赛、DROP、GSM8K、HellaSwag、WinoGrande及涵盖57个学科的MMLU等多种经典自然语言理解与推理任务。它提供了每个任务的细粒度结果,包括准确率、F1分数等指标及其标准误差,是衡量模型在零样本或少样本设置下泛化能力的标准基准,广泛应用于大型语言模型的性能横向比较与能力图谱绘制。
解决学术问题
该数据集有效解决了大语言模型评估缺乏统一、透明、可复现基准的学术困境。通过标准化评测流程与细粒度结果记录,它使得研究者能够客观比较不同模型在数学推理(GSM8K)、常识推理(HellaSwag、WinoGrande)、阅读理解(DROP)以及广泛学科知识(MMLU)上的表现差异。这对于揭示模型的能力边界、诊断特定任务上的弱点、以及推动模型在逻辑推理与知识记忆方面的改进具有奠基性意义。
实际应用
在实际应用中,该评测数据集为模型选型与部署决策提供了量化依据。开发者可以依据模型在各任务上的表现,判断其是否适合用于教育辅导、智能问答、代码生成辅助或专业领域知识检索等场景。例如,GSM8K的数学推理能力关乎自动化教育工具的有效性,而MMLU的多学科表现则决定了模型在科研辅助或跨领域信息整合中的可靠性,从而指导特定业务场景下的模型定制与优化方向。
数据集最近研究
最新研究方向
当前,大语言模型(LLM)的性能评估已成为自然语言处理领域的核心议题,而开放式LLM排行榜(Open LLM Leaderboard)的出现,为模型间的公平比较提供了标准化平台。围绕facebook/opt-iml-max-1.3b这一模型,最新研究方向聚焦于多任务泛化能力的精细化评测,特别是在常识推理(如Winogrande)、数学推理(如GSM8K)以及阅读理解(如DROP)等基准上的表现。该数据集记录了模型在64个配置下的详尽结果,揭示了模型在复杂推理任务上的局限性——例如在GSM8K上准确率仅约0.5%,凸显了当前模型在数学推理领域的瓶颈。这一研究方向与业界对模型鲁棒性和可解释性的迫切需求紧密相连,推动了评估指标从单一准确率向包含F1分数和标准误差的多维度量体系演进。其重要意义在于,通过公开透明的评估数据,为后续模型优化提供了关键基准,并促进了社区对LLM能力的系统性理解,进而引导研究资源向薄弱环节倾斜。
以上内容由遇见数据集搜集并总结生成
二维码
社区交流群
二维码
科研交流群
商业服务