open-llm-leaderboard-old/details_KoboldAI__OPT-13B-Nerybus-Mix

该数据集是在评估模型KoboldAI/OPT-13B-Nerybus-Mix时自动生成的，用于Open LLM Leaderboard的评估。数据集包含64个配置，每个配置对应一个评估任务。数据集由2次运行生成，每次运行的结果存储为特定的split，split的名称使用运行的时间戳。train split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

该数据集是在对模型 KoboldAI/OPT-13B-Nerybus-Mix 进行评估运行期间自动创建的，用于 Open LLM Leaderboard。

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 2 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" 分割总是指向最新的结果。
• 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。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_KoboldAI__OPT-13B-Nerybus-Mix", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-13T14:02:37.898876 运行的最新结果：

python { "all": { "em": 0.001363255033557047, "em_stderr": 0.0003778609196460603, "f1": 0.05258703859060418, "f1_stderr": 0.0012667306875041833, "acc": 0.3412797478178501, "acc_stderr": 0.00740815003235425 }, "harness|drop|3": { "em": 0.001363255033557047, "em_stderr": 0.0003778609196460603, "f1": 0.05258703859060418, "f1_stderr": 0.0012667306875041833 }, "harness|gsm8k|5": { "acc": 0.0037907505686125853, "acc_stderr": 0.0016927007401501795 }, "harness|winogrande|5": { "acc": 0.6787687450670876, "acc_stderr": 0.013123599324558321 } }

配置详情

• harness_arc_challenge_25

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|arc:challenge|25_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|arc:challenge|25_2023-07-19T18:44:38.472954.parquet

• harness_drop_3

  • 分割：2023_10_13T14_02_37.898876
  • 路径：**/details_harness|drop|3_2023-10-13T14-02-37.898876.parquet
  • 分割：latest
  • 路径：**/details_harness|drop|3_2023-10-13T14-02-37.898876.parquet

• harness_gsm8k_5

  • 分割：2023_10_13T14_02_37.898876
  • 路径：**/details_harness|gsm8k|5_2023-10-13T14-02-37.898876.parquet
  • 分割：latest
  • 路径：**/details_harness|gsm8k|5_2023-10-13T14-02-37.898876.parquet

• harness_hellaswag_10

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hellaswag|10_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hellaswag|10_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:44:38.472954.parquet 等 40 个文件
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:44:38.472954.parquet 等 40 个文件

• harness_hendrycksTest_abstract_algebra_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_anatomy_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_astronomy_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_business_ethics_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_clinical_knowledge_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_college_biology_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_college_chemistry_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_college_computer_science_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_college_mathematics_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_college_medicine_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_college_physics_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_computer_security_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_conceptual_physics_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_econometrics_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:44:38.472954.parquet
  • 分割：latest
  • 路径：**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:44:38.472954.parquet

• harness_hendrycksTest_electrical_engineering_5

  • 分割：2023_07_19T18_44_38.472954
  • 路径：**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:44:38.472954.parquet