open-llm-leaderboard-old/details_ajibawa-2023__Uncensored-Jordan-13B

该数据集是在模型ajibawa-2023/Uncensored-Jordan-13B的评估运行期间自动创建的，用于在Open LLM Leaderboard上进行评估。数据集由64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。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，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

该数据集是在评估模型ajibawa-2023/Uncensored-Jordan-13B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64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，每次运行的详细结果可以在每个配置中找到，以运行的时间戳命名的特定分割形式存储。
• "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_ajibawa-2023__Uncensored-Jordan-13B_public", "harness_winogrande_5", split="train")

最新结果

以下是最新结果（注意可能包含其他任务的结果）：

python { "all": { "acc": 0.5551346436733366, "acc_stderr": 0.033773935379363566, "acc_norm": 0.5623156588862028, "acc_norm_stderr": 0.03452935511879212, "mc1": 0.35006119951040393, "mc1_stderr": 0.01669794942015103, "mc2": 0.5051246541228124, "mc2_stderr": 0.015683474268697605, "em": 0.10371224832214765, "em_stderr": 0.003122327158910168, "f1": 0.1647325922818787, "f1_stderr": 0.003269141000174996 }, ... }

配置详情

• config_name: harness_arc_challenge_25

  • split: 2023_11_18T18_01_22.350849
    • path: **/details_harness|arc:challenge|25_2023-11-18T18-01-22.350849.parquet
  • split: latest
    • path: **/details_harness|arc:challenge|25_2023-11-18T18-01-22.350849.parquet

• config_name: harness_drop_3

  • split: 2023_11_18T18_01_22.350849
    • path: **/details_harness|drop|3_2023-11-18T18-01-22.350849.parquet
  • split: latest
    • path: **/details_harness|drop|3_2023-11-18T18-01-22.350849.parquet

• config_name: harness_gsm8k_5

  • split: 2023_11_18T18_01_22.350849
    • path: **/details_harness|gsm8k|5_2023-11-18T18-01-22.350849.parquet
  • split: latest
    • path: **/details_harness|gsm8k|5_2023-11-18T18-01-22.350849.parquet

• config_name: harness_hellaswag_10

  • split: 2023_11_18T18_01_22.350849
    • path: **/details_harness|hellaswag|10_2023-11-18T18-01-22.350849.parquet
  • split: latest
    • path: **/details_harness|hellaswag|10_2023-11-18T18-01-22.350849.parquet

• config_name: harness_hendrycksTest_5

  • split: 2023_11_18T18_01_22.350849
    • path:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-11-18T18-01-22.350849.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-11-18T18-01-22.350849.parquet
      • ... }