open-llm-leaderboard-old/details_FuseAI__OpenChat-3.5-7B-Solar

该数据集是在评估模型FuseAI/OpenChat-3.5-7B-Solar时自动创建的，主要用于在Open LLM Leaderboard上进行评估。数据集包含63个配置，每个配置对应一个评估任务。数据集由1次运行生成，每次运行的结果存储在不同的分割中，分割名称使用运行的时间戳。train分割始终指向最新的结果。此外，还有一个名为results的配置，存储了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。

原始信息汇总

数据集概述

该数据集是在评估模型 FuseAI/OpenChat-3.5-7B-Solar 在 Open LLM Leaderboard 上的运行过程中自动创建的。

数据集组成

• 数据集包含 63 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。
• 额外的 "results" 配置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 2024-02-29T19:13:37.785528 运行的最新结果：

python { "all": { "acc": 0.640989352644531, "acc_stderr": 0.032219950430538366, "acc_norm": 0.6426863775225623, "acc_norm_stderr": 0.03286986373227689, "mc1": 0.30966952264381886, "mc1_stderr": 0.01618574435514491, "mc2": 0.4564538987524744, "mc2_stderr": 0.015087244368040306 }, "harness|arc:challenge|25": { "acc": 0.5844709897610921, "acc_stderr": 0.01440136664121638, "acc_norm": 0.6296928327645052, "acc_norm_stderr": 0.01411129875167495 }, "harness|hellaswag|10": { "acc": 0.6486755626369249, "acc_stderr": 0.004764084597176899, "acc_norm": 0.8418641704839673, "acc_norm_stderr": 0.003641226294167794 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.35, "acc_stderr": 0.0479372485441102, "acc_norm": 0.35, "acc_norm_stderr": 0.0479372485441102 }, # 其他任务的结果... }

配置详情

• config_name: harness_arc_challenge_25

  • split: 2024_02_29T19_13_37.785528
    • path: **/details_harness|arc:challenge|25_2024-02-29T19-13-37.785528.parquet
  • split: latest
    • path: **/details_harness|arc:challenge|25_2024-02-29T19-13-37.785528.parquet

• config_name: harness_gsm8k_5

  • split: 2024_02_29T19_13_37.785528
    • path: **/details_harness|gsm8k|5_2024-02-29T19-13-37.785528.parquet
  • split: latest
    • path: **/details_harness|gsm8k|5_2024-02-29T19-13-37.785528.parquet

• config_name: harness_hellaswag_10

  • split: 2024_02_29T19_13_37.785528
    • path: **/details_harness|hellaswag|10_2024-02-29T19-13-37.785528.parquet
  • split: latest
    • path: **/details_harness|hellaswag|10_2024-02-29T19-13-37.785528.parquet

• config_name: harness_hendrycksTest_5

  • split: 2024_02_29T19_13_37.785528
    • path:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2024-02-29T19-13-37.785528.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2024-02-29T19-13-37.785528.parquet

      其他路径...

}