five

open-llm-leaderboard/details_PulsarAI__Chat-AYB-Nova-13B

收藏
Hugging Face2023-10-27 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_PulsarAI__Chat-AYB-Nova-13B
下载链接
链接失效反馈
官方服务:
资源简介:
--- pretty_name: Evaluation run of PulsarAI/Chat-AYB-Nova-13B dataset_summary: "Dataset automatically created during the evaluation run of model\ \ [PulsarAI/Chat-AYB-Nova-13B](https://huggingface.co/PulsarAI/Chat-AYB-Nova-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_PulsarAI__Chat-AYB-Nova-13B\"\ ,\n\t\"harness_winogrande_5\",\n\tsplit=\"train\")\n```\n\n## Latest results\n\n\ These are the [latest results from run 2023-10-27T20:18:17.450635](https://huggingface.co/datasets/open-llm-leaderboard/details_PulsarAI__Chat-AYB-Nova-13B/blob/main/results_2023-10-27T20-18-17.450635.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\": 0.0041946308724832215,\n\ \ \"em_stderr\": 0.0006618716168266419,\n \"f1\": 0.0802946728187919,\n\ \ \"f1_stderr\": 0.0016873252068220475,\n \"acc\": 0.44971346473405205,\n\ \ \"acc_stderr\": 0.010392725523775513\n },\n \"harness|drop|3\": {\n\ \ \"em\": 0.0041946308724832215,\n \"em_stderr\": 0.0006618716168266419,\n\ \ \"f1\": 0.0802946728187919,\n \"f1_stderr\": 0.0016873252068220475\n\ \ },\n \"harness|gsm8k|5\": {\n \"acc\": 0.12357846853677028,\n \ \ \"acc_stderr\": 0.009065050306776921\n },\n \"harness|winogrande|5\"\ : {\n \"acc\": 0.7758484609313339,\n \"acc_stderr\": 0.011720400740774104\n\ \ }\n}\n```" repo_url: https://huggingface.co/PulsarAI/Chat-AYB-Nova-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_10_08T14_44_32.660445 path: - '**/details_harness|arc:challenge|25_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|arc:challenge|25_2023-10-08T14-44-32.660445.parquet' - config_name: harness_drop_3 data_files: - split: 2023_10_27T20_18_17.450635 path: - '**/details_harness|drop|3_2023-10-27T20-18-17.450635.parquet' - split: latest path: - '**/details_harness|drop|3_2023-10-27T20-18-17.450635.parquet' - config_name: harness_gsm8k_5 data_files: - split: 2023_10_27T20_18_17.450635 path: - '**/details_harness|gsm8k|5_2023-10-27T20-18-17.450635.parquet' - split: latest path: - '**/details_harness|gsm8k|5_2023-10-27T20-18-17.450635.parquet' - config_name: harness_hellaswag_10 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hellaswag|10_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hellaswag|10_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-management|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-anatomy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-astronomy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_biology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-college_physics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-computer_security|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-econometrics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-global_facts|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-human_aging|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-international_law|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-management|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-marketing|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-nutrition|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-philosophy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-prehistory|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_law|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-public_relations|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-security_studies|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-sociology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-virology|5_2023-10-08T14-44-32.660445.parquet' - '**/details_harness|hendrycksTest-world_religions|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_abstract_algebra_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_anatomy_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-anatomy|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-anatomy|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_astronomy_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-astronomy|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-astronomy|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_business_ethics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-business_ethics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_clinical_knowledge_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_college_biology_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-college_biology|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_biology|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_college_chemistry_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_chemistry|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_college_computer_science_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_computer_science|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_college_mathematics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_mathematics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_college_medicine_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_medicine|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_college_physics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-college_physics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-college_physics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_computer_security_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-computer_security|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-computer_security|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_conceptual_physics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_econometrics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-econometrics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-econometrics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_electrical_engineering_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_elementary_mathematics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_formal_logic_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-formal_logic|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_global_facts_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-global_facts|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-global_facts|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_biology_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_biology|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_chemistry_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_computer_science_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_european_history_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_geography_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_geography|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_government_and_politics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_macroeconomics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_mathematics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_microeconomics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_physics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_physics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_psychology_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_statistics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_us_history_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_high_school_world_history_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_human_aging_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-human_aging|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_aging|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_human_sexuality_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-human_sexuality|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_international_law_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-international_law|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-international_law|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_jurisprudence_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-jurisprudence|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_logical_fallacies_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_machine_learning_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-machine_learning|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_management_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-management|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-management|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_marketing_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-marketing|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-marketing|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_medical_genetics_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-medical_genetics|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_miscellaneous_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-miscellaneous|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_moral_disputes_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_disputes|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_moral_scenarios_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_nutrition_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-nutrition|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-nutrition|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_philosophy_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-philosophy|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-philosophy|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_prehistory_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-prehistory|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-prehistory|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_professional_accounting_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_accounting|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_professional_law_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-professional_law|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_law|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_professional_medicine_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_medicine|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_professional_psychology_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-professional_psychology|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_public_relations_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-public_relations|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-public_relations|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_security_studies_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-security_studies|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-security_studies|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_sociology_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-sociology|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-sociology|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_us_foreign_policy_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_virology_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-virology|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-virology|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_hendrycksTest_world_religions_5 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|hendrycksTest-world_religions|5_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|hendrycksTest-world_religions|5_2023-10-08T14-44-32.660445.parquet' - config_name: harness_truthfulqa_mc_0 data_files: - split: 2023_10_08T14_44_32.660445 path: - '**/details_harness|truthfulqa:mc|0_2023-10-08T14-44-32.660445.parquet' - split: latest path: - '**/details_harness|truthfulqa:mc|0_2023-10-08T14-44-32.660445.parquet' - config_name: harness_winogrande_5 data_files: - split: 2023_10_27T20_18_17.450635 path: - '**/details_harness|winogrande|5_2023-10-27T20-18-17.450635.parquet' - split: latest path: - '**/details_harness|winogrande|5_2023-10-27T20-18-17.450635.parquet' - config_name: results data_files: - split: 2023_10_08T14_44_32.660445 path: - results_2023-10-08T14-44-32.660445.parquet - split: 2023_10_27T20_18_17.450635 path: - results_2023-10-27T20-18-17.450635.parquet - split: latest path: - results_2023-10-27T20-18-17.450635.parquet --- # Dataset Card for Evaluation run of PulsarAI/Chat-AYB-Nova-13B ## Dataset Description - **Homepage:** - **Repository:** https://huggingface.co/PulsarAI/Chat-AYB-Nova-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 [PulsarAI/Chat-AYB-Nova-13B](https://huggingface.co/PulsarAI/Chat-AYB-Nova-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_PulsarAI__Chat-AYB-Nova-13B", "harness_winogrande_5", split="train") ``` ## Latest results These are the [latest results from run 2023-10-27T20:18:17.450635](https://huggingface.co/datasets/open-llm-leaderboard/details_PulsarAI__Chat-AYB-Nova-13B/blob/main/results_2023-10-27T20-18-17.450635.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": 0.0041946308724832215, "em_stderr": 0.0006618716168266419, "f1": 0.0802946728187919, "f1_stderr": 0.0016873252068220475, "acc": 0.44971346473405205, "acc_stderr": 0.010392725523775513 }, "harness|drop|3": { "em": 0.0041946308724832215, "em_stderr": 0.0006618716168266419, "f1": 0.0802946728187919, "f1_stderr": 0.0016873252068220475 }, "harness|gsm8k|5": { "acc": 0.12357846853677028, "acc_stderr": 0.009065050306776921 }, "harness|winogrande|5": { "acc": 0.7758484609313339, "acc_stderr": 0.011720400740774104 } } ``` ### Supported Tasks and Leaderboards [More Information Needed] ### Languages [More Information Needed] ## Dataset Structure ### Data Instances [More Information Needed] ### Data Fields [More Information Needed] ### Data Splits [More Information Needed] ## Dataset Creation ### Curation Rationale [More Information Needed] ### Source Data #### Initial Data Collection and Normalization [More Information Needed] #### Who are the source language producers? [More Information Needed] ### Annotations #### Annotation process [More Information Needed] #### Who are the annotators? [More Information Needed] ### Personal and Sensitive Information [More Information Needed] ## Considerations for Using the Data ### Social Impact of Dataset [More Information Needed] ### Discussion of Biases [More Information Needed] ### Other Known Limitations [More Information Needed] ## Additional Information ### Dataset Curators [More Information Needed] ### Licensing Information [More Information Needed] ### Citation Information [More Information Needed] ### Contributions [More Information Needed]
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集简介

该数据集是在模型 PulsarAI/Chat-AYB-Nova-13B 的评估运行期间自动创建的,用于 Open LLM Leaderboard

数据集结构

  • 配置数量:64个配置,每个配置对应一个评估任务。
  • 运行次数:数据集从2次运行中创建。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。
  • 训练分割:"train" 分割始终指向最新的结果。
  • 结果配置: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_PulsarAI__Chat-AYB-Nova-13B", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-27T20:18:17.450635 运行 的最新结果: python { "all": { "em": 0.0041946308724832215, "em_stderr": 0.0006618716168266419, "f1": 0.0802946728187919, "f1_stderr": 0.0016873252068220475, "acc": 0.44971346473405205, "acc_stderr": 0.010392725523775513 }, "harness|drop|3": { "em": 0.0041946308724832215, "em_stderr": 0.0006618716168266419, "f1": 0.0802946728187919, "f1_stderr": 0.0016873252068220475 }, "harness|gsm8k|5": { "acc": 0.12357846853677028, "acc_stderr": 0.009065050306776921 }, "harness|winogrande|5": { "acc": 0.7758484609313339, "acc_stderr": 0.011720400740774104 } }

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_10_08T14_44_32.660445
      • 路径:**/details_harness|arc:challenge|25_2023-10-08T14-44-32.660445.parquet
    • 分割:latest
      • 路径:**/details_harness|arc:challenge|25_2023-10-08T14-44-32.660445.parquet
  • harness_drop_3
    • 分割:2023_10_27T20_18_17.450635
      • 路径:**/details_harness|drop|3_2023-10-27T20-18-17.450635.parquet
    • 分割:latest
      • 路径:**/details_harness|drop|3_2023-10-27T20-18-17.450635.parquet
  • harness_gsm8k_5
    • 分割:2023_10_27T20_18_17.450635
      • 路径:**/details_harness|gsm8k|5_2023-10-27T20-18-17.450635.parquet
    • 分割:latest
      • 路径:**/details_harness|gsm8k|5_2023-10-27T20-18-17.450635.parquet
  • harness_hellaswag_10
    • 分割:2023_10_08T14_44_32.660445
      • 路径:**/details_harness|hellaswag|10_2023-10-08T14-44-32.660445.parquet
    • 分割:latest
      • 路径:**/details_harness|hellaswag|10_2023-10-08T14-44-32.660445.parquet
  • harness_hendrycksTest_5
    • 分割:2023_10_08T14_44_32.660445
      • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-formal_logic|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-global_facts|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_biology|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_geography|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_physics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-human_aging|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-human_sexuality|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-international_law|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-jurisprudence|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-machine_learning|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-management|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-marketing|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-medical_genetics|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-miscellaneous|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-moral_disputes|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-nutrition|5_2023-10-08T14-44-32.660445.parquet
      • 路径:**/details_harness|hendrycksTest-philosophy|5_2023-10-08T14-44-32.660445.parquet
      • 路径:`**/
搜集汇总
数据集介绍
main_image_url
构建方式
该数据集是在Open LLM Leaderboard框架下,针对PulsarAI/Chat-AYB-Nova-13B模型进行自动化评估过程中生成的。它整合了多次运行的结果,每次运行对应一组特定的评估任务,共计64个配置项,每个配置项对应一个独立的任务。数据以Parquet格式存储,并通过时间戳命名的方式区分不同运行批次,其中'train'分割始终指向最新一次评估的产出。
特点
数据集的核心特色在于其多任务、多运行的结构化设计。它涵盖了从常识推理(如Winogrande)到数学求解(如GSM8K)等多样化的评估维度,并提供了详细的指标(如准确率、F1分数)及其标准误差。通过为每个任务配置独立的数据分割,该数据集能够精确追踪模型在特定领域上的表现变化,支持对模型能力的细粒度剖析。
使用方法
用户可通过HuggingFace Datasets库便捷地加载数据。指定配置名称(如'harness_winogrande_5')和分割(如'train')即可获取特定任务的评估详情。例如,使用`load_dataset("open-llm-leaderboard/details_PulsarAI__Chat-AYB-Nova-13B", "harness_winogrande_5", split="train")`来加载最新结果。此外,'results'配置提供了所有任务的聚合指标,便于整体性能对比与可视化分析。
背景与挑战
背景概述
随着大语言模型的蓬勃发展,如何系统、公正地评估模型性能成为学界与工业界共同关注的焦点。在此背景下,HuggingFace社区于2023年推出了Open LLM Leaderboard,旨在为各类开源大模型提供标准化的评测平台。该数据集正是针对PulsarAI团队开发的Chat-AYB-Nova-13B模型在Leaderboard上的评测结果而构建,由HuggingFace研究团队负责创建与管理,核心联系人Clémentine Fourrier主导了数据集的自动化生成流程。数据集覆盖了ARC、DROP、GSM8K、HellaSwag、MMLU及WinoGrande等多项经典基准测试,通过64个配置项详细记录了模型在推理、数学、常识与知识理解等多维度的表现,为社区提供了透明、可复现的模型能力参考,对推动大模型评估体系的规范化具有重要示范意义。
当前挑战
该数据集所应对的核心挑战在于大语言模型评估标准的碎片化与结果不可复现问题。具体而言,不同研究团队常采用差异化的评测协议与提示策略,导致模型性能难以横向比较;同时,传统评测结果多以静态表格呈现,缺乏细粒度的中间过程记录,阻碍了错误模式的深入分析。在数据集构建过程中,挑战主要体现在自动化流水线的工程实现上:需要将模型在多个独立任务上的推理日志、预测结果与评测指标进行结构化整合,并确保不同时间点的运行结果能以统一格式存储,同时维持64个配置项与多个时间戳分片之间的数据一致性。此外,如何高效处理大规模评测产生的Parquet格式文件,并设计直观的接口供研究者按需加载特定任务或运行批次的数据,也是构建过程中的技术难点。
常用场景
经典使用场景
在大型语言模型的评估体系中,Open LLM Leaderboard 的细粒度评测数据集为模型性能的横向对比提供了标准化基准。该数据集记录了 PulsarAI/Chat-AYB-Nova-13B 模型在 ARC-Challenge、HellaSwag、MMLU、GSM8K、Winogrande 和 DROP 等经典任务上的推理结果,覆盖常识推理、数学求解、文本理解和知识问答等核心维度。研究者可通过加载特定任务的配置与分割,复现模型在零样本或少样本设定下的准确率、F1 分数等指标,从而系统性地剖析模型在语言理解、逻辑推理与知识迁移方面的能力边界。
实际应用
在实际应用中,该数据集支撑了模型选型与部署前的效能评估。开发者可依据 GSM8K 的数学推理成绩筛选适用于教育辅导场景的模型,或依据 Winogrande 的指代消解表现优化对话系统的上下文理解能力。企业可复用该数据集的评测流水线,快速验证微调后的模型在垂直领域(如法律文书分析或医学问答)的适配性,降低试错成本。此外,其结构化结果便于集成至自动化模型监控平台,持续跟踪生产环境中模型性能的退化趋势。
衍生相关工作
该数据集衍生了一系列关于评测方法论与模型改进的经典工作。基于其输出的细粒度错误分析,催生了针对特定能力短板(如数学符号推理与常识冲突检测)的增强训练策略。研究者利用该数据集构建了模型能力雷达图,启发多任务学习与知识蒸馏的优化方向。同时,其公开的评测日志为分析指令微调与强化学习对齐技术的边际收益提供了数据支撑,间接推动了 RLHF 与 RLAIF 等对齐算法的迭代。该数据集已成为连接模型开发与能力诊断的核心枢纽。
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