open-llm-leaderboard-old/details_player1537__dolphinette

该数据集是在Open LLM Leaderboard上对模型player1537/dolphinette进行评估时自动生成的。数据集由64个配置组成，每个配置对应一个评估任务。数据集由2次运行生成，每次运行的结果作为特定配置中的一个分割，分割名称使用运行的时间戳。train分割始终指向最新的结果。此外，还有一个名为results的配置，存储了所有运行的聚合结果，并用于在Open LLM Leaderboard上计算和显示聚合指标。

原始信息汇总

数据集概述

该数据集是在对模型 player1537/dolphinette 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。每个运行可以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。

加载数据示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_player1537__dolphinette", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-23T20:59:16.214990 运行 的最新结果： python { "all": { "em": 0.04194630872483222, "em_stderr": 0.0020529645493585565, "f1": 0.0809899328859061, "f1_stderr": 0.002298583656346721, "acc": 0.27111286503551696, "acc_stderr": 0.007001142252211221 }, "harness|drop|3": { "em": 0.04194630872483222, "em_stderr": 0.0020529645493585565, "f1": 0.0809899328859061, "f1_stderr": 0.002298583656346721 }, "harness|gsm8k|5": { "acc": 0.0, "acc_stderr": 0.0 }, "harness|winogrande|5": { "acc": 0.5422257300710339, "acc_stderr": 0.014002284504422442 } }

配置详情

• harness_arc_challenge_25

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|arc:challenge|25_2023-09-16T10-13-56.989865.parquet

• harness_drop_3

  • 分割: 2023_10_23T20_59_16.214990, latest
  • 路径: **/details_harness|drop|3_2023-10-23T20-59-16.214990.parquet

• harness_gsm8k_5

  • 分割: 2023_10_23T20_59_16.214990, latest
  • 路径: **/details_harness|gsm8k|5_2023-10-23T20-59-16.214990.parquet

• harness_hellaswag_10

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hellaswag|10_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: 多个路径，如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T10-13-56.989865.parquet 等

• harness_hendrycksTest_abstract_algebra_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_anatomy_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_astronomy_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_business_ethics_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_clinical_knowledge_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_college_biology_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_college_chemistry_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_college_computer_science_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_college_mathematics_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_college_medicine_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_college_physics_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_computer_security_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_conceptual_physics_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_econometrics_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_electrical_engineering_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-16T10-13-56.989865.parquet

• harness_hendrycksTest_elementary_mathematics_5

  • 分割: 2023_09_16T10_13_56.989865, latest
  • 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-16T10-13-56.989865.parquet