five

open-llm-leaderboard-old/details_vikp__phi2

收藏
Hugging Face2023-10-26 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_vikp__phi2
下载链接
链接失效反馈
官方服务:
资源简介:
--- pretty_name: Evaluation run of vikp/phi2 dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [vikp/phi2](https://huggingface.co/vikp/phi2) 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_vikp__phi2\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-10-26T20:53:51.880388](https://huggingface.co/datasets/open-llm-leaderboard/details_vikp__phi2/blob/main/results_2023-10-26T20-53-51.880388.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\": 0.0005243288590604027,\n\ \ \"em_stderr\": 0.00023443780464835683,\n \"f1\": 0.02841547818791949,\n\ \ \"f1_stderr\": 0.0009543917019203069,\n \"acc\": 0.2634748167903622,\n\ \ \"acc_stderr\": 0.008154316883299411\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.0005243288590604027,\n \"em_stderr\": 0.00023443780464835683,\n\ \ \"f1\": 0.02841547818791949,\n \"f1_stderr\": 0.0009543917019203069\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.006823351023502654,\n \ \ \"acc_stderr\": 0.002267537102254493\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.5201262825572218,\n \"acc_stderr\": 0.014041096664344329\n\ \ }\n}\n```" repo_url: https://huggingface.co/vikp/phi2 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_10_04T03_36_11.040901 path: - '**/details_harness|arc:challenge|25_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-10-04T03-36-11.040901.parquet' - config_name: harness_drop_3 data_files: - split: 2023_10_26T20_53_51.880388 path: - '**/details_harness|drop|3_2023-10-26T20-53-51.880388.parquet' - split: latest path: - '**/details_harness|drop|3_2023-10-26T20-53-51.880388.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_10_26T20_53_51.880388 path: - '**/details_harness|gsm8k|5_2023-10-26T20-53-51.880388.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-10-26T20-53-51.880388.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hellaswag|10_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-management|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-management|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-10-04T03-36-11.040901.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-international_law|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-management|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-marketing|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-sociology|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-virology|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-10-04T03-36-11.040901.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_10_04T03_36_11.040901 path: - '**/details_harness|truthfulqa:mc|0_2023-10-04T03-36-11.040901.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-10-04T03-36-11.040901.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_10_26T20_53_51.880388 path: - '**/details_harness|winogrande|5_2023-10-26T20-53-51.880388.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-10-26T20-53-51.880388.parquet' - config_name: results data_files: - split: 2023_10_04T03_36_11.040901 path: - results_2023-10-04T03-36-11.040901.parquet - split: 2023_10_26T20_53_51.880388 path: - results_2023-10-26T20-53-51.880388.parquet - split: latest path: - results_2023-10-26T20-53-51.880388.parquet --- # Dataset Card for Evaluation run of vikp/phi2 ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/vikp/phi2 - **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 [vikp/phi2](https://huggingface.co/vikp/phi2) 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_vikp__phi2", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-10-26T20:53:51.880388](https://huggingface.co/datasets/open-llm-leaderboard/details_vikp__phi2/blob/main/results_2023-10-26T20-53-51.880388.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": 0.0005243288590604027, "em_stderr": 0.00023443780464835683, "f1": 0.02841547818791949, "f1_stderr": 0.0009543917019203069, "acc": 0.2634748167903622, "acc_stderr": 0.008154316883299411 }, "harness|drop|3": { "em": 0.0005243288590604027, "em_stderr": 0.00023443780464835683, "f1": 0.02841547818791949, "f1_stderr": 0.0009543917019203069 }, "harness|gsm8k|5": { "acc": 0.006823351023502654, "acc_stderr": 0.002267537102254493 }, "harness|winogrande|5": { "acc": 0.5201262825572218, "acc_stderr": 0.014041096664344329 } } ``` ### 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]
原始信息汇总

数据集概述

数据集来源

该数据集是在评估模型 vikp/phi2Open LLM Leaderboard 上的自动创建的。

数据集组成

数据集由 64 个配置组成,每个配置对应一个评估任务。

数据集创建

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。"train" 分割总是指向最新的结果。

额外配置

一个额外的配置 "results" 存储所有运行的聚合结果,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_vikp__phi2", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-26T20:53:51.880388 运行 的最新结果: python { "all": { "em": 0.0005243288590604027, "em_stderr": 0.00023443780464835683, "f1": 0.02841547818791949, "f1_stderr": 0.0009543917019203069, "acc": 0.2634748167903622, "acc_stderr": 0.008154316883299411 }, "harness|drop|3": { "em": 0.0005243288590604027, "em_stderr": 0.00023443780464835683, "f1": 0.02841547818791949, "f1_stderr": 0.0009543917019203069 }, "harness|gsm8k|5": { "acc": 0.006823351023502654, "acc_stderr": 0.002267537102254493 }, "harness|winogrande|5": { "acc": 0.5201262825572218, "acc_stderr": 0.014041096664344329 } }

配置详情

配置列表

  • harness_arc_challenge_25

    • 分割: 2023_10_04T03_36_11.040901
    • 路径: **/details_harness|arc:challenge|25_2023-10-04T03-36-11.040901.parquet
    • 分割: latest
    • 路径: **/details_harness|arc:challenge|25_2023-10-04T03-36-11.040901.parquet
  • harness_drop_3

    • 分割: 2023_10_26T20_53_51.880388
    • 路径: **/details_harness|drop|3_2023-10-26T20-53-51.880388.parquet
    • 分割: latest
    • 路径: **/details_harness|drop|3_2023-10-26T20-53-51.880388.parquet
  • harness_gsm8k_5

    • 分割: 2023_10_26T20_53_51.880388
    • 路径: **/details_harness|gsm8k|5_2023-10-26T20-53-51.880388.parquet
    • 分割: latest
    • 路径: **/details_harness|gsm8k|5_2023-10-26T20-53-51.880388.parquet
  • harness_hellaswag_10

    • 分割: 2023_10_04T03_36_11.040901
    • 路径: **/details_harness|hellaswag|10_2023-10-04T03-36-11.040901.parquet
    • 分割: latest
    • 路径: **/details_harness|hellaswag|10_2023-10-04T03-36-11.040901.parquet
  • harness_hendrycksTest_5

    • 分割: 2023_10_04T03_36_11.040901
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-management|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T03-36-11.040901.parquet
      • **/details_harness|hendrycksTest-professional_law|5_2023-10-04T03-36-11.040901.parquet
      • `**/details_harness|hendrycksTest-professional_medicine|5_20
搜集汇总
数据集介绍
main_image_url
构建方式
在大型语言模型评估领域,对模型性能进行系统化、可复现的评测至关重要。该数据集是在Open LLM Leaderboard框架下,针对模型vikp/phi2进行自动化评测过程中自动生成的。数据集包含64个配置,每个配置对应一项评测任务,例如arc_challenge、drop、gsm8k、hellaswag以及涵盖57个学科的hendrycksTest等。数据来源于两次独立的运行,每次运行的结果以时间戳命名作为数据集中的特定分割,而'train'分割始终指向最新一次运行的结果。此外,一个名为'results'的额外配置存储了所有运行的聚合结果,用于计算并展示Open LLM Leaderboard上的综合指标。
特点
该数据集最为显著的特点在于其结构化的评测记录体系。它完整保留了模型在多项基准任务上的原始评测细节,为后续的深度分析提供了坚实基础。每个配置下的数据均以Parquet格式存储,高效且易于处理。通过为每次运行创建独立的时间戳分割,并设置'train'分割指向最新结果,数据集实现了对模型性能演变的追踪与对比。而'results'配置则提供了便捷的聚合视角,使得研究人员能够快速获取模型在各项任务上的平均表现及其标准误差,从而直观评估模型的综合能力。
使用方法
研究人员可通过HuggingFace的datasets库便捷地加载该数据集。例如,使用load_dataset函数指定数据集名称'open-llm-leaderboard/details_vikp__phi2'、具体评测任务的配置名称(如'harness_winogrande_5')以及所需的分割(如'train'),即可获取该任务的最新评测详情。若要访问历史运行数据,只需将分割名称替换为对应的时间戳字符串。这种设计使得对模型性能进行纵向比较和细粒度分析变得简单直接,极大地便利了自然语言处理领域的研究与开发工作。
背景与挑战
背景概述
大语言模型(LLM)的性能评估是推动自然语言处理领域发展的关键环节。Open LLM Leaderboard由Hugging Face团队(主要联系人为Clémentine Fourrier)于2023年创建,旨在为开源LLM提供标准化、透明化的多任务评测平台。该数据集专为模型vikp/phi2的评测结果而设计,记录了其在ARC-Challenge、HellaSwag、MMLU(涵盖57个学科)、GSM8K、Winogrande及DROP等多样化任务上的表现。通过自动化的评测流水线,数据集不仅存储了每次运行(如2023年10月4日与10月26日)的细粒度指标,还聚合了全局结果,为社区提供了可复现的模型能力基准。这一工作填补了开源LLM评测缺乏统一标准和细粒度细节的空白,显著提升了模型对比的公平性与科研透明度。
当前挑战
该数据集所解决的领域问题在于:开源LLM的评测长期面临任务碎片化、指标不统一及结果难以复现的挑战。具体而言,不同研究团队常采用私有数据集或自定义评估脚本,导致模型间的公平比较极为困难。在构建过程中,数据集面临的技术挑战包括:1)需兼容语言模型评估基准(LM Evaluation Harness)中数十种异构任务(如推理、常识、数学和阅读理解),并统一其评分逻辑;2)需设计高效的数据存储结构,以管理多次运行产生的海量配置(本数据集含64种配置)与时间戳分片;3)需确保评测流水线的自动化与可扩展性,使新模型能无缝接入并自动生成标准化报告。这些挑战的克服,使得该数据集成为LLM评估领域的重要基础设施。
常用场景
经典使用场景
在大型语言模型评估领域,open-llm-leaderboard-old/details_vikp__phi2数据集扮演着不可或缺的基准角色。该数据集专为记录和复现模型vikp/phi2在Open LLM Leaderboard上的评测轨迹而设计,涵盖了从ARC挑战、HellaSwag到GSM8K、Winogrande等64项经典自然语言理解与推理任务。研究者可精准加载各任务配置下的详细评估结果,从而系统性地剖析模型的推理能力、常识知识储备及数学解题水平。这一结构化评测数据为对比不同规模、不同架构的语言模型提供了标准化参照,成为检验模型综合性能的黄金标尺。
实际应用
在实际应用中,该数据集为工业界部署语言模型提供了严谨的准入测试框架。企业在将phi2等模型集成至智能客服、教育辅导或代码生成系统前,可借助此数据集模拟多场景压力测试,精准量化模型在数学推理、常识问答及文本蕴含等任务上的可靠性。例如,GSM8K任务的准确率直接关联模型解决实际数学问题的能力,而Winogrande的评测结果则反映模型处理代词指代歧义的鲁棒性。这种基于标准化评测的决策流程,显著降低了模型上线后的风险,加速了从研究原型到商业产品的转化进程。
衍生相关工作
基于此数据集衍生的经典工作主要集中在评估方法论的创新与模型性能的深度解析。研究者利用其细粒度的任务配置,开发了诸如任务难度加权聚合、置信区间分析等高级评估工具,进一步提升了评测的科学性。同时,该数据集催生了多篇关于模型能力边界探索的论文,例如通过分析phi2在不同时间戳评测中的性能波动,揭示了训练数据时效性对模型知识更新的影响。这些衍生工作不仅丰富了语言模型评估的理论体系,也为后续构建更全面的评测基准(如涵盖多模态或长文本任务)奠定了坚实基础。
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