遇见数据集

open-llm-leaderboard/details_HWERI__pythia-1.4b-deduped-sharegpt

收藏
Hugging Face2023-09-16 更新2024-03-04 收录
官方服务:

资源简介:

--- pretty_name: Evaluation run of HWERI/pythia-1.4b-deduped-sharegpt dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [HWERI/pythia-1.4b-deduped-sharegpt](https://huggingface.co/HWERI/pythia-1.4b-deduped-sharegpt)\ \ 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_HWERI__pythia-1.4b-deduped-sharegpt\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-09-16T20:15:27.580598](https://huggingface.co/datasets/open-llm-leaderboard/details_HWERI__pythia-1.4b-deduped-sharegpt/blob/main/results_2023-09-16T20-15-27.580598.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\": 0.0010486577181208054,\n\ \ \"em_stderr\": 0.00033145814652192217,\n \"f1\": 0.04875104865771823,\n\ \ \"f1_stderr\": 0.0012458540332815637,\n \"acc\": 0.2804129195481258,\n\ \ \"acc_stderr\": 0.008239894933698364\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.0010486577181208054,\n \"em_stderr\": 0.00033145814652192217,\n\ \ \"f1\": 0.04875104865771823,\n \"f1_stderr\": 0.0012458540332815637\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.008339651250947688,\n \ \ \"acc_stderr\": 0.002504942226860534\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.5524861878453039,\n \"acc_stderr\": 0.013974847640536194\n\ \ }\n}\n```" repo_url: https://huggingface.co/HWERI/pythia-1.4b-deduped-sharegpt 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_17T18_24_42.073512 path: - '**/details_harness|arc:challenge|25_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-08-17T18:24:42.073512.parquet' - config_name: harness_drop_3 data_files: - split: 2023_09_16T20_15_27.580598 path: - '**/details_harness|drop|3_2023-09-16T20-15-27.580598.parquet' - split: latest path: - '**/details_harness|drop|3_2023-09-16T20-15-27.580598.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_09_16T20_15_27.580598 path: - '**/details_harness|gsm8k|5_2023-09-16T20-15-27.580598.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-09-16T20-15-27.580598.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hellaswag|10_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-management|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-management|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-08-17T18:24:42.073512.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-international_law|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-management|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-marketing|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-sociology|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-virology|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-08-17T18:24:42.073512.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_08_17T18_24_42.073512 path: - '**/details_harness|truthfulqa:mc|0_2023-08-17T18:24:42.073512.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-08-17T18:24:42.073512.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_09_16T20_15_27.580598 path: - '**/details_harness|winogrande|5_2023-09-16T20-15-27.580598.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-09-16T20-15-27.580598.parquet' - config_name: results data_files: - split: 2023_08_17T18_24_42.073512 path: - results_2023-08-17T18:24:42.073512.parquet - split: 2023_09_16T20_15_27.580598 path: - results_2023-09-16T20-15-27.580598.parquet - split: latest path: - results_2023-09-16T20-15-27.580598.parquet --- # Dataset Card for Evaluation run of HWERI/pythia-1.4b-deduped-sharegpt ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/HWERI/pythia-1.4b-deduped-sharegpt - **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 [HWERI/pythia-1.4b-deduped-sharegpt](https://huggingface.co/HWERI/pythia-1.4b-deduped-sharegpt) 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_HWERI__pythia-1.4b-deduped-sharegpt", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-09-16T20:15:27.580598](https://huggingface.co/datasets/open-llm-leaderboard/details_HWERI__pythia-1.4b-deduped-sharegpt/blob/main/results_2023-09-16T20-15-27.580598.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": 0.0010486577181208054, "em_stderr": 0.00033145814652192217, "f1": 0.04875104865771823, "f1_stderr": 0.0012458540332815637, "acc": 0.2804129195481258, "acc_stderr": 0.008239894933698364 }, "harness|drop|3": { "em": 0.0010486577181208054, "em_stderr": 0.00033145814652192217, "f1": 0.04875104865771823, "f1_stderr": 0.0012458540332815637 }, "harness|gsm8k|5": { "acc": 0.008339651250947688, "acc_stderr": 0.002504942226860534 }, "harness|winogrande|5": { "acc": 0.5524861878453039, "acc_stderr": 0.013974847640536194 } } ``` ### 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 HWERI/pythia-1.4b-deduped-sharegpt 在 Open LLM Leaderboard 上的自动创建的。

数据集结构

  • 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作为一个特定的分割找到,分割名称使用运行的时间戳。
  • "train" 分割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_HWERI__pythia-1.4b-deduped-sharegpt", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-16T20:15:27.580598 运行的最新结果: python { "all": { "em": 0.0010486577181208054, "em_stderr": 0.00033145814652192217, "f1": 0.04875104865771823, "f1_stderr": 0.0012458540332815637, "acc": 0.2804129195481258, "acc_stderr": 0.008239894933698364 }, "harness|drop|3": { "em": 0.0010486577181208054, "em_stderr": 0.00033145814652192217, "f1": 0.04875104865771823, "f1_stderr": 0.0012458540332815637 }, "harness|gsm8k|5": { "acc": 0.008339651250947688, "acc_stderr": 0.002504942226860534 }, "harness|winogrande|5": { "acc": 0.5524861878453039, "acc_stderr": 0.013974847640536194 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|arc:challenge|25_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|arc:challenge|25_2023-08-17T18:24:42.073512.parquet
  • harness_drop_3

    • 分割:2023_09_16T20_15_27.580598
    • 路径:**/details_harness|drop|3_2023-09-16T20-15-27.580598.parquet
    • 分割:latest
    • 路径:**/details_harness|drop|3_2023-09-16T20-15-27.580598.parquet
  • harness_gsm8k_5

    • 分割:2023_09_16T20_15_27.580598
    • 路径:**/details_harness|gsm8k|5_2023-09-16T20-15-27.580598.parquet
    • 分割:latest
    • 路径:**/details_harness|gsm8k|5_2023-09-16T20-15-27.580598.parquet
  • harness_hellaswag_10

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hellaswag|10_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hellaswag|10_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet 等 50 个文件
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet 等 50 个文件
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:24:42.073512.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-08-17T18:24:42.073512.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_08_17T18_24_42.073512
    • 路径:`**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T18:24:42.07
搜集汇总
数据集介绍
open-llm-leaderboard/details_HWERI__pythia-1.4b-deduped-sharegpt 数据集图片
构建方式
该数据集源自Open LLM Leaderboard平台对模型HWERI/pythia-1.4b-deduped-sharegpt的自动评估过程。构建方式上,数据集由64个配置组成,每个配置对应于一项被评估的任务。评估共执行了两次运行,每次运行的结果被存储为对应配置下的一个独立分割(split),分割名称采用运行的时间戳进行标识。同时,所有配置中均包含一个名为“train”的分割,该分割始终指向最新一次运行的评估结果。此外,数据集还额外提供了一个名为“results”的配置,用于汇总并存储所有运行的聚合指标,这些聚合指标被用于在Open LLM Leaderboard上计算和展示模型的综合表现。
使用方法
使用该数据集时,研究者可通过Hugging Face的datasets库进行加载。具体而言,调用load_dataset函数并指定数据集名称,同时传入目标任务的配置名称(如"harness_winogrande_5")以及所需的分割名称(如"train"以获取最新结果,或使用时间戳字符串获取特定历史运行结果)。加载后返回的数据对象可直接用于分析模型在各任务上的细粒度表现。对于需要整体概览的场景,访问"results"配置即可获取所有评估任务的聚合指标,便于进行模型间的横向比较与性能基准的建立。
背景与挑战
背景概述
在大型语言模型(LLM)性能评估领域,Open LLM Leaderboard由Hugging Face团队于2023年推出,旨在为日益增长的开源语言模型提供一个标准化、透明化的评测平台。该数据集记录了模型HWERI/pythia-1.4b-deduped-sharegpt在2023年8月至9月期间的两次评估运行结果,覆盖了ARC挑战、DROP、GSM8K、HellaSwag及涵盖57个学科的MMLU等多项核心基准任务。作为Pythia系列模型的变体,该模型通过ShareGPT数据微调,其评估结果不仅反映了特定训练策略对1.4B参数规模模型性能的影响,也为社区理解模型在推理、常识与数学等维度上的能力边界提供了实证依据,推动了开源LLM评估体系的规范化发展。
当前挑战
该数据集所反映的核心挑战体现在多个层面。在领域问题层面,LLM评估面临任务多样性与泛化能力的矛盾——模型在Winogrande上取得55.25%的准确率,却在GSM8K数学推理任务中仅达0.83%,凸显了当前模型在复杂推理与常识理解之间的显著能力鸿沟。在构建过程中,数据集需整合来自不同时间戳的多次运行结果,并以64个配置项分别对应各评估任务,这带来了数据版本管理与结果聚合的复杂性,尤其是当后续评估未覆盖相同任务时,如何确保结果的可比性与一致性成为关键难题。
常用场景
经典使用场景
在大型语言模型评估领域,该数据集被广泛用于对模型进行多维度、标准化的性能基准测试。其核心使用方式是通过加载特定任务配置(如winogrande、gsm8k、drop等),获取模型在各项指标上的详细表现,从而实现对模型推理、常识理解、数学计算等能力的系统化度量。研究者常利用此数据集对pythia-1.4b-deduped-sharegpt等模型进行纵向或横向比较,以验证模型优化策略的有效性。
解决学术问题
该数据集有效解决了大语言模型评估中缺乏统一、可复现基准的学术难题。通过提供结构化、多任务、多轮次的评估结果存储,它使得研究者能够精确量化模型在推理、常识、数学等核心能力上的表现差异,并分析不同训练策略(如去重、微调)对模型性能的影响。其意义在于为模型对比提供了标准化平台,推动了评估流程的透明化和可复现性,对理解模型能力边界与改进方向具有重要学术价值。
实际应用
在实际应用中,该数据集及其背后的评估体系被广泛用于模型选型与部署决策。企业或研究机构在挑选适合特定场景(如教育问答、客服系统、内容生成)的语言模型时,可依据该数据集提供的多任务评分,快速判断模型在推理、常识、数学等方面的适配性。此外,该数据集的持续更新机制使得用户能够追踪模型迭代过程中的性能变化,为模型上线前的质量把控提供了数据支撑。
数据集最近研究
最新研究方向
当前,大语言模型评估体系正逐步从单一基准测试向多维能力图谱演进,这一变革通过Open LLM Leaderboard等标准化平台得以具象化。本数据集记录了pythia-1.4b-deduped-sharegpt模型在包括ARC挑战、DROP、GSM8K、WinoGrande及涵盖57个学科领域的Hendrycks测试等多元任务上的细致表现,揭示了小规模模型在常识推理与数学问题求解中的能力边界。尤其值得关注的是,该模型在WinoGrande上达到55.25%的准确率,展现了去重训练数据与指令微调对缓解歧义理解的协同作用,而在GSM8K上不足1%的准确率则凸显了算术推理仍是轻量级模型的攻坚难点。这一评估范式不仅为模型迭代提供了可复现的量化基准,更推动研究者重新审视数据质量与任务难度之间的微妙平衡,对构建更具鲁棒性的评估协议具有启示意义。
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