five

open-llm-leaderboard/details_facebook__galactica-1.3b

收藏
Hugging Face2023-09-02 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_facebook__galactica-1.3b
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在模型 facebook/galactica-1.3b 在 Open LLM Leaderboard 上的评估运行期间自动创建的。数据集由 61 个配置组成,每个配置对应一个评估任务。数据集是从 1 次运行中创建的,每次运行在每个配置中作为一个特定的分割找到,分割名称使用运行的时间戳。train 分割始终指向最新的结果。一个额外的配置 results 存储了运行的所有聚合结果,用于计算和显示 Open LLM Leaderboard 上的聚合指标。文件还提供了如何使用 Python 代码加载运行细节的示例,并列出了特定运行的最新结果。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集摘要

该数据集是在对模型 facebook/galactica-1.3b 进行评估运行时自动创建的,用于 Open LLM Leaderboard

数据集组成

  • 该数据集包含 61 个配置,每个配置对应一个评估任务。
  • 数据集从 1 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。

最新结果

以下是 2023-09-02T13:58:08.758268 运行的最新结果:

python { "all": { "acc": 0.2724936919306173, "acc_stderr": 0.03226048262400512, "acc_norm": 0.2744720231892299, "acc_norm_stderr": 0.03227238640653428, "mc1": 0.2484700122399021, "mc1_stderr": 0.015127427096520667, "mc2": 0.41399712836660274, "mc2_stderr": 0.01494063292915903 }, "harness|arc:challenge|25": { "acc": 0.2960750853242321, "acc_stderr": 0.013340916085246261, "acc_norm": 0.3412969283276451, "acc_norm_stderr": 0.01385583128749772 }, "harness|hellaswag|10": { "acc": 0.3375821549492133, "acc_stderr": 0.004719187890948067, "acc_norm": 0.40908185620394344, "acc_norm_stderr": 0.004906595857916765 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.26, "acc_stderr": 0.0440844002276808, "acc_norm": 0.26, "acc_norm_stderr": 0.0440844002276808 }, # 其他任务的结果省略... }

配置详情

  • config_name: harness_arc_challenge_25

    • 分割: 2023_09_02T13_58_08.758268
      • 路径: **/details_harness|arc:challenge|25_2023-09-02T13:58:08.758268.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2023-09-02T13:58:08.758268.parquet
  • config_name: harness_hellaswag_10

    • 分割: 2023_09_02T13_58_08.758268
      • 路径: **/details_harness|hellaswag|10_2023-09-02T13:58:08.758268.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2023-09-02T13:58:08.758268.parquet
  • config_name: harness_hendrycksTest_5

    • 分割: 2023_09_02T13_58_08.758268
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-management|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-professional_law|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-professional_medicine|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-professional_psychology|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-public_relations|5_2023-09-02T13:58:08.758268.parquet
        • **/details_harness|hendrycksTest-security_studies|5_2023-09-02T13:58:08.758268.parquet
        • `**/details_harness|hendrycksTest-
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作