open-llm-leaderboard-old/details_chargoddard__servile-harpsichord-cdpo

该数据集是在模型chargoddard/servile-harpsichord-cdpo在Open LLM Leaderboard上进行评估时自动创建的。数据集由63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，每次运行都可以在特定配置中找到，分割名称使用运行的时间戳命名。train分割始终指向最新结果。此外，results配置存储了所有运行的聚合结果，并用于计算和显示Open LLM Leaderboard上的聚合指标。

原始信息汇总

数据集概述

数据集来源

该数据集是在评估模型chargoddard/servile-harpsichord-cdpo在Open LLM Leaderboard上的运行过程中自动创建的。

数据集组成

数据集由63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，每次运行可以在每个配置中找到特定的拆分，拆分名称使用运行的时间戳。"train"拆分始终指向最新的结果。

数据集加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_chargoddard__servile-harpsichord-cdpo", "harness_winogrande_5", split="train")

最新结果

这些是最新结果，来自2023-12-10T06:44:09.091422的运行。每个任务的结果可以在"results"配置和每个评估的"latest"拆分中找到。

配置详情

以下是部分配置及其数据文件路径：

• config_name: harness_arc_challenge_25

  • split: 2023_12_10T06_44_09.091422
    • path: **/details_harness|arc:challenge|25_2023-12-10T06-44-09.091422.parquet
  • split: latest
    • path: **/details_harness|arc:challenge|25_2023-12-10T06-44-09.091422.parquet

• config_name: harness_gsm8k_5

  • split: 2023_12_10T06_44_09.091422
    • path: **/details_harness|gsm8k|5_2023-12-10T06-44-09.091422.parquet
  • split: latest
    • path: **/details_harness|gsm8k|5_2023-12-10T06-44-09.091422.parquet

• config_name: harness_hellaswag_10

  • split: 2023_12_10T06_44_09.091422
    • path: **/details_harness|hellaswag|10_2023-12-10T06-44-09.091422.parquet
  • split: latest
    • path: **/details_harness|hellaswag|10_2023-12-10T06-44-09.091422.parquet

• config_name: harness_hendrycksTest_5

  • split: 2023_12_10T06_44_09.091422
    • path:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-12-10T06-44-09.091422.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-12-10T06-44-09.091422.parquet