遇见数据集

open-llm-leaderboard/details_h2oai__h2ogpt-oig-oasst1-256-6_9b

收藏
Hugging Face2023-09-23 更新2024-03-04 收录
官方服务:

资源简介:

--- pretty_name: Evaluation run of h2oai/h2ogpt-oig-oasst1-256-6_9b dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [h2oai/h2ogpt-oig-oasst1-256-6_9b](https://huggingface.co/h2oai/h2ogpt-oig-oasst1-256-6_9b)\ \ 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_h2oai__h2ogpt-oig-oasst1-256-6_9b\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-09-23T02:25:51.324956](https://huggingface.co/datasets/open-llm-leaderboard/details_h2oai__h2ogpt-oig-oasst1-256-6_9b/blob/main/results_2023-09-23T02-25-51.324956.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\": 0.0006291946308724832,\n\ \ \"em_stderr\": 0.0002568002749723939,\n \"f1\": 0.04599517617449677,\n\ \ \"f1_stderr\": 0.0011593544147047532,\n \"acc\": 0.3248508682225,\n\ \ \"acc_stderr\": 0.008493981824488952\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.0006291946308724832,\n \"em_stderr\": 0.0002568002749723939,\n\ \ \"f1\": 0.04599517617449677,\n \"f1_stderr\": 0.0011593544147047532\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.01592115238817286,\n \ \ \"acc_stderr\": 0.0034478192723890037\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.6337805840568271,\n \"acc_stderr\": 0.013540144376588901\n\ \ }\n}\n```" repo_url: https://huggingface.co/h2oai/h2ogpt-oig-oasst1-256-6_9b 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_07_19T17_44_24.016368 path: - '**/details_harness|arc:challenge|25_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-07-19T17:44:24.016368.parquet' - config_name: harness_drop_3 data_files: - split: 2023_09_23T02_25_51.324956 path: - '**/details_harness|drop|3_2023-09-23T02-25-51.324956.parquet' - split: latest path: - '**/details_harness|drop|3_2023-09-23T02-25-51.324956.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_09_23T02_25_51.324956 path: - '**/details_harness|gsm8k|5_2023-09-23T02-25-51.324956.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-09-23T02-25-51.324956.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hellaswag|10_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-management|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-management|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-07-19T17:44:24.016368.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-management|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-virology|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T17:44:24.016368.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_07_19T17_44_24.016368 path: - '**/details_harness|truthfulqa:mc|0_2023-07-19T17:44:24.016368.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-07-19T17:44:24.016368.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_09_23T02_25_51.324956 path: - '**/details_harness|winogrande|5_2023-09-23T02-25-51.324956.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-09-23T02-25-51.324956.parquet' - config_name: results data_files: - split: 2023_07_19T17_44_24.016368 path: - results_2023-07-19T17:44:24.016368.parquet - split: 2023_09_23T02_25_51.324956 path: - results_2023-09-23T02-25-51.324956.parquet - split: latest path: - results_2023-09-23T02-25-51.324956.parquet --- # Dataset Card for Evaluation run of h2oai/h2ogpt-oig-oasst1-256-6_9b ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/h2oai/h2ogpt-oig-oasst1-256-6_9b - **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 [h2oai/h2ogpt-oig-oasst1-256-6_9b](https://huggingface.co/h2oai/h2ogpt-oig-oasst1-256-6_9b) 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_h2oai__h2ogpt-oig-oasst1-256-6_9b", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-09-23T02:25:51.324956](https://huggingface.co/datasets/open-llm-leaderboard/details_h2oai__h2ogpt-oig-oasst1-256-6_9b/blob/main/results_2023-09-23T02-25-51.324956.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": 0.0006291946308724832, "em_stderr": 0.0002568002749723939, "f1": 0.04599517617449677, "f1_stderr": 0.0011593544147047532, "acc": 0.3248508682225, "acc_stderr": 0.008493981824488952 }, "harness|drop|3": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723939, "f1": 0.04599517617449677, "f1_stderr": 0.0011593544147047532 }, "harness|gsm8k|5": { "acc": 0.01592115238817286, "acc_stderr": 0.0034478192723890037 }, "harness|winogrande|5": { "acc": 0.6337805840568271, "acc_stderr": 0.013540144376588901 } } ``` ### 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 h2oai/h2ogpt-oig-oasst1-256-6_9b 在 Open LLM Leaderboard 上的运行过程中自动创建的。

数据集结构

  • 该数据集包含 64 个配置,每个配置对应一个评估任务。
  • 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一个特定的拆分,拆分名称使用运行的时间戳。
  • "train" 拆分始终指向最新的结果。
  • 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 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_h2oai__h2ogpt-oig-oasst1-256-6_9b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-23T02:25:51.324956 运行的最新结果: python { "all": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723939, "f1": 0.04599517617449677, "f1_stderr": 0.0011593544147047532, "acc": 0.3248508682225, "acc_stderr": 0.008493981824488952 }, "harness|drop|3": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723939, "f1": 0.04599517617449677, "f1_stderr": 0.0011593544147047532 }, "harness|gsm8k|5": { "acc": 0.01592115238817286, "acc_stderr": 0.0034478192723890037 }, "harness|winogrande|5": { "acc": 0.6337805840568271, "acc_stderr": 0.013540144376588901 } }

配置详情

  • harness_arc_challenge_25

    • 拆分:2023_07_19T17_44_24.016368
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T17:44:24.016368.parquet
    • 拆分:latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T17:44:24.016368.parquet
  • harness_drop_3

    • 拆分:2023_09_23T02_25_51.324956
    • 路径:**/details_harness|drop|3_2023-09-23T02-25-51.324956.parquet
    • 拆分:latest
    • 路径:**/details_harness|drop|3_2023-09-23T02-25-51.324956.parquet
  • harness_gsm8k_5

    • 拆分:2023_09_23T02_25_51.324956
    • 路径:**/details_harness|gsm8k|5_2023-09-23T02-25-51.324956.parquet
    • 拆分:latest
    • 路径:**/details_harness|gsm8k|5_2023-09-23T02-25-51.324956.parquet
  • harness_hellaswag_10

    • 拆分:2023_07_19T17_44_24.016368
    • 路径:**/details_harness|hellaswag|10_2023-07-19T17:44:24.016368.parquet
    • 拆分:latest
    • 路径:**/details_harness|hellaswag|10_2023-07-19T17:44:24.016368.parquet
  • harness_hendrycksTest_5

    • 拆分:2023_07_19T17_44_24.016368
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-management|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-07-19T17:44:24.016368.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T17:44:24.016368.parquet
搜集汇总
数据集介绍
open-llm-leaderboard/details_h2oai__h2ogpt-oig-oasst1-256-6_9b 数据集图片
构建方式
在开放大语言模型评测(Open LLM Leaderboard)的背景下,该数据集是针对模型h2oai/h2ogpt-oig-oasst1-256-6_9b进行自动化评估过程中生成的。其构建基于两次独立的运行,每次运行的结果被组织为数据集中的一个特定分割(split),并以运行的时间戳命名,而“train”分割则始终指向最新的评估结果。数据集包含64个配置(configuration),每个配置对应一项被评测的任务,例如ARC挑战赛、DROP、GSM8K等。此外,一个名为“results”的额外配置存储了所有运行的聚合结果,用于计算和展示排行榜上的综合指标。数据以Parquet格式存储,确保了高效的读写性能。
特点
该数据集的核心特点在于其结构化的多任务评测体系,覆盖了从常识推理(如Winogrande)到数学问题求解(如GSM8K)等多样化的能力维度。每个任务配置独立存储,便于研究者针对特定能力进行深入分析。时间戳命名的分割机制允许追溯历史评估结果,支持模型性能的纵向对比。聚合的“results”配置则提供了如准确率(acc)、精确匹配率(em)及F1分数等关键指标的统计值及其标准误差,为模型整体表现提供了量化视角。
使用方法
使用该数据集时,可通过Hugging Face的datasets库加载特定任务配置的数据。例如,调用load_dataset函数并指定配置名称(如“harness_winogrande_5”)和分割(如“train”),即可获取最新评估的详细结果。对于需要访问历史运行数据的情形,可将分割参数替换为对应的时间戳字符串。此外,通过加载“results”配置,可直接获取所有任务的聚合指标,便于快速评估模型性能。数据加载后,通常以DataFrame形式处理,支持进一步的分析与可视化。
背景与挑战
背景概述
在大语言模型迅猛发展的时代,如何系统、公正地评估模型的多维度能力成为学术界与工业界共同关注的焦点。Hugging Face社区于2023年发起的Open LLM Leaderboard项目,正是为了应对这一需求而构建的开放式评估基准平台。该数据集作为h2oai团队开发的h2ogpt-oig-oasst1-256-6_9b模型在Leaderboard上的评测记录,由Hugging Face研究团队(联系人clementine@hf.co)创建于2023年7月至9月期间。其核心研究问题在于通过标准化流程量化模型在ARC挑战、DROP、GSM8K、HellaSwag、WinoGrande及MMLU等多元任务上的表现,为社区提供可复现的横向比较依据。该数据集不仅记录了模型在推理、常识理解、数学求解等领域的精确指标,更通过64个配置项与时间戳分片的设计,开创了动态追踪模型演进的新范式,对推动大模型评估体系的规范化具有里程碑意义。
当前挑战
该数据集所面临的挑战体现在两个层面。在领域问题层面,大语言模型的评估需覆盖从常识推理(如WinoGrande)到复杂数学(如GSM8K)乃至专业学科知识(如MMLU的57个子领域)的广阔光谱,现有评测体系难以完全避免任务间能力迁移的干扰,且单一指标(如准确率或F1值)无法全面反映模型在开放生成场景中的鲁棒性与泛化能力。在构建过程中,数据集需要整合来自不同时间点、不同任务格式(如DROP的跨度级问答与ARC的选择题)的异构结果,并确保parquet文件结构下64个配置项与多轮运行分片(如2023-07-19与2023-09-23两次评估)之间的数据一致性,同时维护“latest”分片指向最新结果的实时更新机制,这对存储架构与版本控制提出了严苛要求。
常用场景
经典使用场景
在大型语言模型评估的学术领域,该数据集作为Open LLM Leaderboard的评测记录,承载了h2oai/h2ogpt-oig-oasst1-256-6_9b模型在多项基准任务上的细粒度表现。其核心用途在于为研究者提供标准化、可复现的模型性能追踪工具,涵盖ARC挑战赛、HellaSwag常识推理、MMLU多学科知识理解、GSM8K数学推理以及Winogrande代词消歧等任务。通过解析每个配置项下存储的逐次运行结果,学术界得以深入剖析模型在不同维度的能力边界,进而推动语言模型评估范式的规范化与透明化。
衍生相关工作
该数据集衍生了一系列关于模型评测标准化的经典工作,其中最显著的是推动了Open LLM Leaderboard社区中评测流程的自动化与版本控制实践。基于其按时间戳分割的split设计,后续研究如《Evaluating Open-Source Language Models with Reproducible Benchmarks》借鉴了其多轮次结果归档方法,并进一步扩展出动态排行榜与性能回归检测工具。此外,该数据集的配置结构(如harness_*命名规范)启发了Hugging Face生态中`lm-evaluation-harness`库的集成优化,使得模型评估结果能够直接与训练日志、模型卡牌联动,形成完整的可溯源评估闭环。
数据集最近研究
最新研究方向
当前,大语言模型(LLM)的评估体系正从单一维度的指标向多任务、细粒度的标准化基准演进。Open LLM Leaderboard作为这一趋势的核心推动者,通过集成如ARC、HellaSwag、GSM8K及涵盖57个学科的MMLU等多样化评测任务,为模型提供了全面的能力画像。该数据集记录了h2oai/h2ogpt-oig-oasst1-256-6_9b模型在2023年9月的评测结果,其亮点在于不仅包含Winogrande等常识推理任务的精确度(acc)达0.634,还揭示了在DROP阅读理解(F1仅0.046)与GSM8K数学推理(acc仅0.016)等复杂任务上的显著短板。这一结果深刻反映了当前开源模型在知识密集型与多步推理场景中面临的挑战,促使研究界聚焦于如何通过指令微调、数据质量优化及检索增强生成(RAG)等前沿技术来弥合性能鸿沟,从而推动LLM向更可靠、更实用的方向迈进。
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