遇见数据集

open-llm-leaderboard/details_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu

收藏
Hugging Face2023-10-27 更新2024-03-04 收录
官方服务:

资源简介:

--- pretty_name: Evaluation run of NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu](https://huggingface.co/NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu)\ \ 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_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-10-27T20:24:24.496556](https://huggingface.co/datasets/open-llm-leaderboard/details_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu/blob/main/results_2023-10-27T20-24-24.496556.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\": 0.0017827181208053692,\n\ \ \"em_stderr\": 0.00043200973460388544,\n \"f1\": 0.06179635067114092,\n\ \ \"f1_stderr\": 0.0013805757246305051,\n \"acc\": 0.4386580563472483,\n\ \ \"acc_stderr\": 0.010104640124560183\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.0017827181208053692,\n \"em_stderr\": 0.00043200973460388544,\n\ \ \"f1\": 0.06179635067114092,\n \"f1_stderr\": 0.0013805757246305051\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.10462471569370735,\n \ \ \"acc_stderr\": 0.00843066808202928\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.7726913970007893,\n \"acc_stderr\": 0.011778612167091087\n\ \ }\n}\n```" repo_url: https://huggingface.co/NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu 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_09_16T11_41_48.010953 path: - '**/details_harness|arc:challenge|25_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-09-16T11-41-48.010953.parquet' - config_name: harness_drop_3 data_files: - split: 2023_10_27T20_24_24.496556 path: - '**/details_harness|drop|3_2023-10-27T20-24-24.496556.parquet' - split: latest path: - '**/details_harness|drop|3_2023-10-27T20-24-24.496556.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_10_27T20_24_24.496556 path: - '**/details_harness|gsm8k|5_2023-10-27T20-24-24.496556.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-10-27T20-24-24.496556.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hellaswag|10_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-management|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-management|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-09-16T11-41-48.010953.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-international_law|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-management|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-marketing|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-sociology|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-virology|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-09-16T11-41-48.010953.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_09_16T11_41_48.010953 path: - '**/details_harness|truthfulqa:mc|0_2023-09-16T11-41-48.010953.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-09-16T11-41-48.010953.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_10_27T20_24_24.496556 path: - '**/details_harness|winogrande|5_2023-10-27T20-24-24.496556.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-10-27T20-24-24.496556.parquet' - config_name: results data_files: - split: 2023_09_16T11_41_48.010953 path: - results_2023-09-16T11-41-48.010953.parquet - split: 2023_10_27T20_24_24.496556 path: - results_2023-10-27T20-24-24.496556.parquet - split: latest path: - results_2023-10-27T20-24-24.496556.parquet --- # Dataset Card for Evaluation run of NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu - **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 [NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu](https://huggingface.co/NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu) 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_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-10-27T20:24:24.496556](https://huggingface.co/datasets/open-llm-leaderboard/details_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu/blob/main/results_2023-10-27T20-24-24.496556.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": 0.0017827181208053692, "em_stderr": 0.00043200973460388544, "f1": 0.06179635067114092, "f1_stderr": 0.0013805757246305051, "acc": 0.4386580563472483, "acc_stderr": 0.010104640124560183 }, "harness|drop|3": { "em": 0.0017827181208053692, "em_stderr": 0.00043200973460388544, "f1": 0.06179635067114092, "f1_stderr": 0.0013805757246305051 }, "harness|gsm8k|5": { "acc": 0.10462471569370735, "acc_stderr": 0.00843066808202928 }, "harness|winogrande|5": { "acc": 0.7726913970007893, "acc_stderr": 0.011778612167091087 } } ``` ### 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

该数据集是在对模型 NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。

数据集组成

  • 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特定的分割,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_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-27T20:24:24.496556 运行的最新结果:

python { "all": { "em": 0.0017827181208053692, "em_stderr": 0.00043200973460388544, "f1": 0.06179635067114092, "f1_stderr": 0.0013805757246305051, "acc": 0.4386580563472483, "acc_stderr": 0.010104640124560183 }, "harness|drop|3": { "em": 0.0017827181208053692, "em_stderr": 0.00043200973460388544, "f1": 0.06179635067114092, "f1_stderr": 0.0013805757246305051 }, "harness|gsm8k|5": { "acc": 0.10462471569370735, "acc_stderr": 0.00843066808202928 }, "harness|winogrande|5": { "acc": 0.7726913970007893, "acc_stderr": 0.011778612167091087 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_09_16T11_41_48.010953
    • 路径:**/details_harness|arc:challenge|25_2023-09-16T11-41-48.010953.parquet
    • 分割:latest
    • 路径:**/details_harness|arc:challenge|25_2023-09-16T11-41-48.010953.parquet
  • harness_drop_3

    • 分割:2023_10_27T20_24_24.496556
    • 路径:**/details_harness|drop|3_2023-10-27T20-24-24.496556.parquet
    • 分割:latest
    • 路径:**/details_harness|drop|3_2023-10-27T20-24-24.496556.parquet
  • harness_gsm8k_5

    • 分割:2023_10_27T20_24_24.496556
    • 路径:**/details_harness|gsm8k|5_2023-10-27T20-24-24.496556.parquet
    • 分割:latest
    • 路径:**/details_harness|gsm8k|5_2023-10-27T20-24-24.496556.parquet
  • harness_hellaswag_10

    • 分割:2023_09_16T11_41_48.010953
    • 路径:**/details_harness|hellaswag|10_2023-09-16T11-41-48.010953.parquet
    • 分割:latest
    • 路径:**/details_harness|hellaswag|10_2023-09-16T11-41-48.010953.parquet
  • harness_hendrycksTest_5

    • 分割:2023_09_16T11_41_48.010953
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-management|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-09-16T11-41-48.010953.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-09-16T11-41-48.010953.parquet
      • `**/details_harness|hendrycksTest-professional_accounting|5_2023-09-16T11
搜集汇总
数据集介绍
open-llm-leaderboard/details_NekoPunchBBB__Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu 数据集图片
构建方式
在开放大语言模型排行榜(Open LLM Leaderboard)的评估框架下,该数据集基于对NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu模型的两次独立运行而自动生成。数据集由64个配置组成,每个配置对应一项被评估的任务,并额外包含一个名为“results”的配置用于存储所有聚合结果。每次运行的评估细节被保存为独立的数据划分,划分名称以该次运行的时间戳命名,而“train”划分则始终指向最新的评估结果。数据以Parquet格式存储,便于高效读取和分析。
特点
该数据集的核心特色在于其结构化的多层次设计,能够完整记录模型在多项任务上的细粒度评估结果。每个配置对应一个具体任务(如ARC挑战、DROP、GSM8K等),并支持通过时间戳划分回溯历史运行数据。聚合结果配置提供了宏观性能指标,如准确率、F1分数和精确匹配率及其标准误差,使得研究者既能纵览全局性能,又能深入分析单项任务的细节,为模型迭代与比较提供了坚实的数据基础。
使用方法
用户可通过HuggingFace的datasets库便捷地加载该数据集。例如,使用`load_dataset`函数并指定配置名称(如“harness_winogrande_5”)及划分(如“train”),即可获取最新评估结果。若需访问特定历史运行的数据,可将划分参数替换为相应的时间戳字符串。此外,通过加载“results”配置,可以获取所有任务的聚合指标,从而快速完成模型性能的横向对比与趋势分析。
背景与挑战
背景概述
随着大语言模型(LLM)技术的飞速发展,如何客观、全面地评估模型性能成为学术界与工业界共同关注的焦点。Open LLM Leaderboard应运而生,由Hugging Face团队于2023年创建,旨在为社区提供一个标准化、可复现的模型评测平台。该数据集记录了NekoPunchBBB团队基于Llama-2-13b-hf模型,采用Open-Platypus数据集并通过QLoRA多GPU微调后的评估结果。核心研究问题在于探究参数高效微调方法(如QLoRA)在保持模型推理能力的同时,能否在多样化的自然语言理解与推理任务中达到与全参数微调相媲美的性能。该数据集作为Leaderboard的组成部分,为后续研究者提供了宝贵的基准参考,推动了开源LLM生态的透明化与规范化。
当前挑战
当前数据集面临多重挑战。首先,在领域问题层面,LLM的评估需要覆盖广泛的任务类型,包括常识推理(如Winogrande)、数学推理(如GSM8K)、阅读理解(如DROP)以及多学科知识(如MMLU),这要求模型具备泛化能力而非单一任务的过拟合。其次,构建过程中,QLoRA微调策略引入了超参数敏感性,例如低秩矩阵的秩与量化位宽的选择直接影响模型收敛与最终性能,而多GPU并行训练又带来了通信开销与负载均衡的难题。此外,评估结果的时间戳表明模型性能可能随运行批次波动,如何确保评测的稳定性和可复现性仍是亟待解决的问题。
常用场景
经典使用场景
该数据集作为Open LLM Leaderboard评估流程的产物,专门用于记录和存储NekoPunchBBB/Llama-2-13b-hf_Open-Platypus-QLoRA-multigpu模型在多个标准化基准任务上的细粒度评测结果。其经典使用场景涵盖了对大语言模型在ARC挑战、DROP推理、GSM8K数学、HellaSwag常识推理、Winogrande指代消解以及涵盖57个学科的MMLU等任务上的性能追踪。研究者可通过加载不同配置下的parquet文件,精准复现模型在特定任务上的表现,从而实现对模型能力的多维度剖析与横向对比。
解决学术问题
在自然语言处理领域,大语言模型的性能评估长期面临标准不一、复现困难等挑战。该数据集通过结构化的评测记录体系,系统性地解决了模型在推理、知识、数学等多项能力上的可重复评估问题,为学术界提供了统一、透明的基准测试框架。其意义在于推动了模型评估的标准化进程,使得不同微调策略(如QLoRA)和基础架构(如Llama-2-13b)的效能差异得以量化呈现,进而促进了对参数高效微调方法有效性的深入理解。
衍生相关工作
该数据集衍生了一系列关于大语言模型评测方法论与微调策略的经典工作。基于其提供的评测轨迹,研究者得以深入分析QLoRA低秩适配方法在不同任务上的泛化特性,并推动了诸如模型合并、动态少样本学习等后续技术的探索。此外,该数据集的公开记录形式也启发了多版本评测追踪工具的开发,为构建持续性的模型能力演化图谱提供了基础设施,成为连接模型训练与性能验证的关键桥梁。
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