遇见数据集

open-llm-leaderboard/details_Undi95__UndiMix-v1-13b

收藏
Hugging Face2023-10-16 更新2024-03-04 收录
官方服务:

资源简介:

--- pretty_name: Evaluation run of Undi95/UndiMix-v1-13b dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [Undi95/UndiMix-v1-13b](https://huggingface.co/Undi95/UndiMix-v1-13b) 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_Undi95__UndiMix-v1-13b\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-10-16T16:31:03.720074](https://huggingface.co/datasets/open-llm-leaderboard/details_Undi95__UndiMix-v1-13b/blob/main/results_2023-10-16T16-31-03.720074.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\": 0.2600671140939597,\n\ \ \"em_stderr\": 0.004492401208347132,\n \"f1\": 0.34945260067114264,\n\ \ \"f1_stderr\": 0.004422869896423944,\n \"acc\": 0.42730704720576507,\n\ \ \"acc_stderr\": 0.010180773732934644\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.2600671140939597,\n \"em_stderr\": 0.004492401208347132,\n\ \ \"f1\": 0.34945260067114264,\n \"f1_stderr\": 0.004422869896423944\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.10007581501137225,\n \ \ \"acc_stderr\": 0.008266274528685646\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.7545382794001578,\n \"acc_stderr\": 0.012095272937183644\n\ \ }\n}\n```" repo_url: https://huggingface.co/Undi95/UndiMix-v1-13b 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_01T00_10_45.842963 path: - '**/details_harness|arc:challenge|25_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-09-01T00:10:45.842963.parquet' - config_name: harness_drop_3 data_files: - split: 2023_10_16T16_31_03.720074 path: - '**/details_harness|drop|3_2023-10-16T16-31-03.720074.parquet' - split: latest path: - '**/details_harness|drop|3_2023-10-16T16-31-03.720074.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_10_16T16_31_03.720074 path: - '**/details_harness|gsm8k|5_2023-10-16T16-31-03.720074.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-10-16T16-31-03.720074.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hellaswag|10_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-management|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-management|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-09-01T00:10:45.842963.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-international_law|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-management|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-marketing|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-sociology|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-virology|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-09-01T00:10:45.842963.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_09_01T00_10_45.842963 path: - '**/details_harness|truthfulqa:mc|0_2023-09-01T00:10:45.842963.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-09-01T00:10:45.842963.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_10_16T16_31_03.720074 path: - '**/details_harness|winogrande|5_2023-10-16T16-31-03.720074.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-10-16T16-31-03.720074.parquet' - config_name: results data_files: - split: 2023_09_01T00_10_45.842963 path: - results_2023-09-01T00:10:45.842963.parquet - split: 2023_10_16T16_31_03.720074 path: - results_2023-10-16T16-31-03.720074.parquet - split: latest path: - results_2023-10-16T16-31-03.720074.parquet --- # Dataset Card for Evaluation run of Undi95/UndiMix-v1-13b ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/Undi95/UndiMix-v1-13b - **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 [Undi95/UndiMix-v1-13b](https://huggingface.co/Undi95/UndiMix-v1-13b) 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_Undi95__UndiMix-v1-13b", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-10-16T16:31:03.720074](https://huggingface.co/datasets/open-llm-leaderboard/details_Undi95__UndiMix-v1-13b/blob/main/results_2023-10-16T16-31-03.720074.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": 0.2600671140939597, "em_stderr": 0.004492401208347132, "f1": 0.34945260067114264, "f1_stderr": 0.004422869896423944, "acc": 0.42730704720576507, "acc_stderr": 0.010180773732934644 }, "harness|drop|3": { "em": 0.2600671140939597, "em_stderr": 0.004492401208347132, "f1": 0.34945260067114264, "f1_stderr": 0.004422869896423944 }, "harness|gsm8k|5": { "acc": 0.10007581501137225, "acc_stderr": 0.008266274528685646 }, "harness|winogrande|5": { "acc": 0.7545382794001578, "acc_stderr": 0.012095272937183644 } } ``` ### 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

数据集来源

该数据集是在评估模型 Undi95/UndiMix-v1-13bOpen 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找到特定的分割,分割名称使用运行的时间戳。
  • "train" 分割总是指向最新的结果。
  • 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_Undi95__UndiMix-v1-13b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-16T16:31:03.720074 运行 的最新结果: python { "all": { "em": 0.2600671140939597, "em_stderr": 0.004492401208347132, "f1": 0.34945260067114264, "f1_stderr": 0.004422869896423944, "acc": 0.42730704720576507, "acc_stderr": 0.010180773732934644 }, "harness|drop|3": { "em": 0.2600671140939597, "em_stderr": 0.004492401208347132, "f1": 0.34945260067114264, "f1_stderr": 0.004422869896423944 }, "harness|gsm8k|5": { "acc": 0.10007581501137225, "acc_stderr": 0.008266274528685646 }, "harness|winogrande|5": { "acc": 0.7545382794001578, "acc_stderr": 0.012095272937183644 } }

配置详情

  • config_name: harness_arc_challenge_25

    • split: 2023_09_01T00_10_45.842963
      • path: **/details_harness|arc:challenge|25_2023-09-01T00:10:45.842963.parquet
    • split: latest
      • path: **/details_harness|arc:challenge|25_2023-09-01T00:10:45.842963.parquet
  • config_name: harness_drop_3

    • split: 2023_10_16T16_31_03.720074
      • path: **/details_harness|drop|3_2023-10-16T16-31-03.720074.parquet
    • split: latest
      • path: **/details_harness|drop|3_2023-10-16T16-31-03.720074.parquet
  • config_name: harness_gsm8k_5

    • split: 2023_10_16T16_31_03.720074
      • path: **/details_harness|gsm8k|5_2023-10-16T16-31-03.720074.parquet
    • split: latest
      • path: **/details_harness|gsm8k|5_2023-10-16T16-31-03.720074.parquet
  • config_name: harness_hellaswag_10

    • split: 2023_09_01T00_10_45.842963
      • path: **/details_harness|hellaswag|10_2023-09-01T00:10:45.842963.parquet
    • split: latest
      • path: **/details_harness|hellaswag|10_2023-09-01T00:10:45.842963.parquet
  • config_name: harness_hendrycksTest_5

    • split: 2023_09_01T00_10_45.842963
      • path:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-management|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|hendrycksTest-professional_law|5_2023-09-01T00:10:45.842963.parquet
        • **/details_harness|
搜集汇总
数据集介绍
open-llm-leaderboard/details_Undi95__UndiMix-v1-13b 数据集图片
构建方式
在大型语言模型评估领域,模型性能的量化分析是推动技术进步的关键环节。该数据集是Open LLM Leaderboard在评估Undi95/UndiMix-v1-13b模型过程中自动生成的产物,其构建过程紧密围绕评估流程展开。数据集包含64个配置项,每个配置对应一项被评估的任务,例如ARC挑战、DROP、GSM8K、WinoGrande及HellaSwag等。数据来源于两次独立的评估运行,每次运行的结果以时间戳命名的分割形式存储在对应配置中,其中“train”分割始终指向最新一次的评估结果。此外,还设有专门的“results”配置,用于汇聚所有运行的聚合指标,这些指标最终被用于Leaderboard上的综合排名展示。数据以Parquet格式存储,确保了高效读写与可扩展性。
特点
该数据集的核心特点在于其结构化与可追溯性。64个配置项覆盖了从常识推理到数学解题的多样化任务,为研究者提供了细粒度的模型能力剖析。每个配置下的多个分割(如基于时间戳的版本)允许用户回溯模型在不同时间点的表现,从而支持纵向比较与性能趋势分析。数据集的“train”分割自动指向最新结果,简化了实时获取最新评估数据的流程。同时,“results”配置集中存储了如准确率(acc)、精确匹配(em)及F1分数等关键指标,并附带标准误差(stderr),为模型性能的统计显著性检验提供了基础。这种设计兼顾了易用性与深度分析的灵活性,使研究者能够高效地提取洞察。
使用方法
研究者可通过Hugging Face的datasets库便捷地加载该数据集。具体而言,使用load_dataset函数时需指定数据集名称及目标配置,例如加载WinoGrande任务的最新结果:data = load_dataset('open-llm-leaderboard/details_Undi95__UndiMix-v1-13b', 'harness_winogrande_5', split='train')。若需访问特定历史运行的数据,可将分割参数替换为对应的时间戳字符串。此外,加载“results”配置可获取所有任务的聚合指标,便于快速进行模型性能的横向对比。数据集以Parquet文件格式提供,支持高效的列式访问与内存映射,适合在Python环境中结合Pandas等工具进行深度分析。这种标准化的加载方式降低了使用门槛,使研究者能专注于模型评估与改进。
背景与挑战
背景概述
在大型语言模型(LLM)蓬勃发展的时代,如何客观、全面地评估模型性能成为推动技术进步的关键一环。由HuggingFace团队于2023年创建并维护的Open LLM Leaderboard,正是为应对这一需求而生的权威评测平台。该数据集作为对模型Undi95/UndiMix-v1-13b的自动化评估记录,由Clementine(clementine@hf.co)等研究人员主导,旨在通过标准化流程衡量13B参数级别模型的综合能力。其核心研究问题在于,如何通过多维度、多任务的评测体系,揭示模型在常识推理(如Winogrande)、数学求解(如GSM8K)、阅读理解(如DROP)以及广泛学科知识(如MMLU)上的真实表现。该数据集的发布,不仅为开发者提供了模型性能的透明化参考,更推动了LLM评测范式的规范化,对社区内模型迭代与比较产生了深远影响。
当前挑战
该数据集所承载的挑战具有双重维度。在领域问题层面,它直面LLM评估中多任务泛化能力的核心难题,例如在GSM8K任务上模型准确率仅约10%,揭示了当前模型在复杂数学推理上的显著短板;而在DROP阅读理解任务中,F1分数不足35%,凸显了模型对细节信息提取与数值推理的薄弱环节。在构建过程中,挑战则体现为评测流程的复杂性与可复现性:数据集需整合来自不同时间戳(如2023年9月与10月)的多次运行结果,并确保各配置(如64个评估任务)的数据格式统一,同时维护“latest”分片以动态指向最新成绩,这对数据版本管理与自动化管道的稳健性提出了严苛要求。
常用场景
经典使用场景
在大型语言模型飞速演进的浪潮中,如何公正、全面地评估模型性能成为学界与工业界的核心关切。该数据集作为Open LLM Leaderboard对Undi95/UndiMix-v1-13b模型的自动化评测产物,其最经典的使用场景在于为研究者提供一套标准化的多任务评估框架。通过集成ARC-Challenge、HellaSwag、MMLU(涵盖57个学科)、GSM8K、Winogrande及DROP等多样化基准,该数据集允许用户对模型的常识推理、科学知识、数学解题、指代消解及阅读理解等能力进行细粒度剖析,从而精准定位模型在特定认知维度上的优势与短板。
衍生相关工作
该数据集衍生出的经典工作主要围绕评测基准的扩展与模型能力的深度解析展开。其结构化评测数据直接催生了诸如LLM-Perf、EvalPlus等自动化分析工具,这些工具能够从海量评测记录中挖掘模型在不同能力维度的衰减规律。同时,基于该数据集的对比分析,研究者提出了模型能力正交分解理论,将通用能力与任务特异性能力解耦,进而指导了混合专家模型(MoE)架构的优化方向。此外,该数据集的时间序列特性(包含多次评测记录)为研究模型微调过程中的能力漂移现象提供了珍贵素材。
数据集最近研究
最新研究方向
在大型语言模型评估领域,Open LLM Leaderboard 已成为衡量模型综合性能的权威基准平台。针对 Undi95/UndiMix-v1-13b 这一混合架构模型,前沿研究聚焦于其在多任务场景下的泛化能力与鲁棒性。该数据集通过整合 ARC、HellaSwag、MMLU 等数十项挑战性任务,系统性地揭示了模型在常识推理、数学求解与专业知识应用中的表现。当前热点围绕模型在 GSM8K 数学问题与 Winogrande 指代消解上的精度差异展开,其中 75.45% 的指代消解准确率与 10.01% 的数学解题率形成鲜明对比,凸显了混合模型在逻辑推理与数值计算间的能力鸿沟。这一发现不仅为模型优化提供了精准的改进方向,更推动了面向复杂推理的评估体系革新,对构建可信赖的通用人工智能具有里程碑意义。
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