遇见数据集

open-llm-leaderboard-old/details_OpenAssistant__oasst-sft-1-pythia-12b

收藏
Hugging Face2023-10-22 更新2024-06-22 收录
官方服务:

资源简介:

--- pretty_name: Evaluation run of OpenAssistant/oasst-sft-1-pythia-12b dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [OpenAssistant/oasst-sft-1-pythia-12b](https://huggingface.co/OpenAssistant/oasst-sft-1-pythia-12b)\ \ 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_OpenAssistant__oasst-sft-1-pythia-12b\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-10-22T03:38:38.139466](https://huggingface.co/datasets/open-llm-leaderboard/details_OpenAssistant__oasst-sft-1-pythia-12b/blob/main/results_2023-10-22T03-38-38.139466.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\": 0.007340604026845637,\n\ \ \"em_stderr\": 0.000874189687534619,\n \"f1\": 0.06295302013422834,\n\ \ \"f1_stderr\": 0.0015822681875736675,\n \"acc\": 0.3140033976135325,\n\ \ \"acc_stderr\": 0.007883417880991842\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.007340604026845637,\n \"em_stderr\": 0.000874189687534619,\n\ \ \"f1\": 0.06295302013422834,\n \"f1_stderr\": 0.0015822681875736675\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.006065200909780136,\n \ \ \"acc_stderr\": 0.002138670301460446\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.6219415943172849,\n \"acc_stderr\": 0.013628165460523237\n\ \ }\n}\n```" repo_url: https://huggingface.co/OpenAssistant/oasst-sft-1-pythia-12b 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_19T18_16_49.631586 path: - '**/details_harness|arc:challenge|25_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-07-19T18:16:49.631586.parquet' - config_name: harness_drop_3 data_files: - split: 2023_10_22T03_38_38.139466 path: - '**/details_harness|drop|3_2023-10-22T03-38-38.139466.parquet' - split: latest path: - '**/details_harness|drop|3_2023-10-22T03-38-38.139466.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_10_22T03_38_38.139466 path: - '**/details_harness|gsm8k|5_2023-10-22T03-38-38.139466.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-10-22T03-38-38.139466.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hellaswag|10_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-management|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-management|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-07-19T18:16:49.631586.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-management|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-virology|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-07-19T18:16:49.631586.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_07_19T18_16_49.631586 path: - '**/details_harness|truthfulqa:mc|0_2023-07-19T18:16:49.631586.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-07-19T18:16:49.631586.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_10_22T03_38_38.139466 path: - '**/details_harness|winogrande|5_2023-10-22T03-38-38.139466.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-10-22T03-38-38.139466.parquet' - config_name: results data_files: - split: 2023_07_19T18_16_49.631586 path: - results_2023-07-19T18:16:49.631586.parquet - split: 2023_10_22T03_38_38.139466 path: - results_2023-10-22T03-38-38.139466.parquet - split: latest path: - results_2023-10-22T03-38-38.139466.parquet --- # Dataset Card for Evaluation run of OpenAssistant/oasst-sft-1-pythia-12b ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/OpenAssistant/oasst-sft-1-pythia-12b - **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 [OpenAssistant/oasst-sft-1-pythia-12b](https://huggingface.co/OpenAssistant/oasst-sft-1-pythia-12b) 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_OpenAssistant__oasst-sft-1-pythia-12b", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-10-22T03:38:38.139466](https://huggingface.co/datasets/open-llm-leaderboard/details_OpenAssistant__oasst-sft-1-pythia-12b/blob/main/results_2023-10-22T03-38-38.139466.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": 0.007340604026845637, "em_stderr": 0.000874189687534619, "f1": 0.06295302013422834, "f1_stderr": 0.0015822681875736675, "acc": 0.3140033976135325, "acc_stderr": 0.007883417880991842 }, "harness|drop|3": { "em": 0.007340604026845637, "em_stderr": 0.000874189687534619, "f1": 0.06295302013422834, "f1_stderr": 0.0015822681875736675 }, "harness|gsm8k|5": { "acc": 0.006065200909780136, "acc_stderr": 0.002138670301460446 }, "harness|winogrande|5": { "acc": 0.6219415943172849, "acc_stderr": 0.013628165460523237 } } ``` ### 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

数据集简介

该数据集是在对模型 OpenAssistant/oasst-sft-1-pythia-12b 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。

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。
  • 数据集由 2 次运行创建,每次运行在每个配置中作为一个特定的分割存在,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 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_OpenAssistant__oasst-sft-1-pythia-12b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-22T03:38:38.139466 运行的最新结果: python { "all": { "em": 0.007340604026845637, "em_stderr": 0.000874189687534619, "f1": 0.06295302013422834, "f1_stderr": 0.0015822681875736675, "acc": 0.3140033976135325, "acc_stderr": 0.007883417880991842 }, "harness|drop|3": { "em": 0.007340604026845637, "em_stderr": 0.000874189687534619, "f1": 0.06295302013422834, "f1_stderr": 0.0015822681875736675 }, "harness|gsm8k|5": { "acc": 0.006065200909780136, "acc_stderr": 0.002138670301460446 }, "harness|winogrande|5": { "acc": 0.6219415943172849, "acc_stderr": 0.013628165460523237 } }

配置详情

以下是部分配置及其数据文件路径:

  • harness_arc_challenge_25
    • 分割: 2023_07_19T18_16_49.631586
      • 路径: **/details_harness|arc:challenge|25_2023-07-19T18:16:49.631586.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2023-07-19T18:16:49.631586.parquet
  • harness_drop_3
    • 分割: 2023_10_22T03_38_38.139466
      • 路径: **/details_harness|drop|3_2023-10-22T03-38-38.139466.parquet
    • 分割: latest
      • 路径: **/details_harness|drop|3_2023-10-22T03-38-38.139466.parquet
  • harness_gsm8k_5
    • 分割: 2023_10_22T03_38_38.139466
      • 路径: **/details_harness|gsm8k|5_2023-10-22T03-38-38.139466.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2023-10-22T03-38-38.139466.parquet
  • harness_hellaswag_10
    • 分割: 2023_07_19T18_16_49.631586
      • 路径: **/details_harness|hellaswag|10_2023-07-19T18:16:49.631586.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2023-07-19T18:16:49.631586.parquet
  • harness_hendrycksTest_5
    • 分割: 2023_07_19T18_16_49.631586
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-management|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-07-19T18:16:49.631586.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-07-19T18:16:49.631586.parquet
        • `**/details_harness|hendrycksTest-professional_law|5_2023-07-19T18
搜集汇总
数据集介绍
open-llm-leaderboard-old/details_OpenAssistant__oasst-sft-1-pythia-12b 数据集图片
构建方式
在大型语言模型评估的宏观背景下,该数据集作为Open LLM Leaderboard评测流程的产物而诞生,旨在系统化记录模型OpenAssistant/oasst-sft-1-pythia-12b在多项基准任务中的表现。数据集的构建依托于HuggingFace的评估框架,通过自动化流程从两次独立的评测运行中采集原始细节数据。每个评测任务对应一个独立的配置项,共计64个配置,覆盖了如ARC挑战赛、DROP、GSM8K、HellaSwag以及涵盖57个学科的HendrycksTest等多样化任务。每次运行的结果被组织为特定的数据分割,以时间戳命名,而'train'分割始终指向最新的评测结果。此外,增设了一个名为'results'的配置,用于聚合所有运行的总体指标,为排行榜上的最终度量计算提供支撑。所有数据均以Parquet格式存储,确保了高效的数据存取与序列化。
使用方法
对于希望利用该数据集进行深入分析的研究者,推荐通过HuggingFace的datasets库进行高效加载。具体而言,可通过load_dataset函数指定数据集名称、目标配置名称(例如'harness_winogrande_5')以及所需的分割(如'train'或具体的时间戳分割),从而获取对应任务在特定运行下的详细评测记录。加载后的数据可直接用于复现排行榜上的性能指标,或进行自定义的统计分析与可视化。若需获取所有任务的聚合结果,可加载'results'配置,该配置提供了模型在所有基准上的综合表现。此外,通过遍历不同的时间戳分割,用户可以构建模型性能的时间序列,从而评估模型迭代或环境变化带来的影响,为模型优化提供数据驱动的决策支持。
背景与挑战
背景概述
在大规模语言模型迅猛发展的背景下,如何系统性地评估模型在多元任务上的表现成为了一个核心议题。Open LLM Leaderboard由HuggingFace团队于2023年创建,旨在为开源语言模型提供标准化、透明化的评测平台。该数据集作为Leaderboard的一部分,专门记录了OpenAssistant团队开发的oasst-sft-1-pythia-12b模型在多项基准任务上的评估结果,涵盖ARC挑战赛、DROP、GSM8K、HellaSwag及MMLU等广泛领域。其核心研究问题在于通过多维度、多轮次的评测,揭示模型在推理、常识理解与数学能力上的真实水平,为社区提供可复现的性能参考。该数据集对推动开源模型的可比性与进步具有重要影响,已成为衡量模型进展的标杆之一。
当前挑战
该数据集所面临的挑战首先体现在领域问题的复杂性上:语言模型评测需覆盖从常识推理到专业知识的广泛任务,而单一模型往往在特定领域表现不均,例如oasst-sft-1-pythia-12b在DROP任务上的F1分数仅约0.063,暴露出模型在复杂文本理解与精确提取上的显著短板。其次,构建过程中的挑战不容忽视:数据集需整合来自不同时间点、不同任务配置的多次评测运行结果,确保数据格式统一、版本可追溯,并处理如GSM8K仅0.6%准确率等极端低分情况下的数据可靠性问题。此外,评测结果的标准化聚合与实时更新机制也增加了工程实现的难度。
常用场景
经典使用场景
在大型语言模型评估领域,该数据集作为Open LLM Leaderboard的标准化评测工具,被广泛应用于对OpenAssistant/oasst-sft-1-pythia-12b等模型的多维度性能量化。它整合了ARC挑战、DROP、GSM8K、HellaSwag及MMLU(涵盖57个学科)等经典基准任务,为研究者提供细粒度的准确率、F1分数等指标,从而系统性地衡量模型在常识推理、数学计算、阅读理解及专业知识掌握等方面的能力。
解决学术问题
该数据集有效解决了大语言模型评测中缺乏统一、可复现对比基准的学术难题。通过构建包含64个配置的标准化评估框架,它使得不同模型在相同任务上的性能比较成为可能,消除了实验环境差异带来的偏差。其意义在于推动模型优化方向的可量化追踪,为研究社区提供了识别模型短板(如低GSM8K准确率)的客观依据,进而促进针对性改进策略的提出。
实际应用
实际应用中,该数据集为模型部署前的质量检验提供关键参考。开发者可利用其多任务评测结果,判断模型在客服对话(依赖常识推理)、教育辅导(涉及数学与专业知识)等场景的适用性。例如,GSM8K任务的低得分警示模型在数学问题求解上的局限性,而Winogrande的高准确率则暗示其在代词消歧任务上的潜力,从而指导领域微调与模型选型决策。
数据集最近研究
最新研究方向
当前大语言模型评估领域的前沿研究聚焦于构建标准化、多维度的评测框架,以系统性地衡量模型在推理、常识及专业领域知识等维度的能力。OpenAssistant/oasst-sft-1-pythia-12b模型在Open LLM Leaderboard上的评测数据,正是这一趋势的典型代表。该数据集通过整合ARC、HellaSwag、MMLU等多样化任务,揭示了模型在复杂推理(如DROP任务中F1值仅6.3%)与数学求解(GSM8K准确率0.6%)方面仍存在显著短板,而在常识推理(Winogrande准确率62.2%)上表现相对稳健。这些细粒度的评测结果不仅为模型迭代提供了明确方向,更推动了社区对开源模型能力边界的深入认知,其影响已超越单一模型,成为衡量整个领域进展的标尺。
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