open-llm-leaderboard/details_Fredithefish__CrimsonPajama

该数据集是在Open LLM Leaderboard上对模型Fredithefish/CrimsonPajama进行评估时自动创建的。它由64个配置组成，每个配置对应一个被评估的任务。数据集由2次运行生成，每次运行在每个配置中表示为特定的分割。train分割始终指向最新的结果。一个名为results的额外配置存储了所有运行的聚合结果，这些结果用于计算和显示Open LLM Leaderboard上的聚合指标。可以使用`datasets`库中的`load_dataset`函数加载该数据集。This dataset was automatically created during the evaluation of the model Fredithefish/CrimsonPajama on the Open LLM Leaderboard. It consists of 64 configurations, each corresponding to one evaluation task. The dataset is generated from two runs, where each run is represented as a specific split under each configuration. The 'train' split always points to the most recent results. An additional configuration named 'results' stores the aggregated results across all runs, which are used to calculate and display the aggregate metrics on the Open LLM Leaderboard. This dataset can be loaded using the `load_dataset` function from the `datasets` library.

原始信息汇总

数据集概述

该数据集是在评估模型 Fredithefish/CrimsonPajama 在 Open LLM Leaderboard 上的自动创建的。数据集包含 64 个配置，每个配置对应一个评估任务。

数据集结构

• 配置数量：64 个配置
• 数据来源：从 2 次运行中创建，每个运行对应一个特定的时间戳
• 数据分割：每个配置包含多个分割，其中 "train" 分割指向最新的结果
• 结果汇总：一个额外的配置 "results" 存储所有运行的汇总结果，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_Fredithefish__CrimsonPajama", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-17T20:55:57.055960 运行 的最新结果：

python { "all": { "em": 0.006396812080536913, "em_stderr": 0.0008164468837432291, "f1": 0.08161598154362382, "f1_stderr": 0.0017802453361789499, "acc": 0.3286203762267581, "acc_stderr": 0.007694655126017044 }, "harness|drop|3": { "em": 0.006396812080536913, "em_stderr": 0.0008164468837432291, "f1": 0.08161598154362382, "f1_stderr": 0.0017802453361789499 }, "harness|gsm8k|5": { "acc": 0.00530705079605762, "acc_stderr": 0.002001305720948034 }, "harness|winogrande|5": { "acc": 0.6519337016574586, "acc_stderr": 0.013388004531086054 } }

配置详情

以下是部分配置的详细信息：

• harness_arc_challenge_25

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|arc:challenge|25_2023-07-19T19:19:26.317110.parquet

• harness_drop_3

  • 分割：2023_10_17T20_55_57.055960, latest
  • 路径：**/details_harness|drop|3_2023-10-17T20-55-57.055960.parquet

• harness_gsm8k_5

  • 分割：2023_10_17T20_55_57.055960, latest
  • 路径：**/details_harness|gsm8k|5_2023-10-17T20-55-57.055960.parquet

• harness_hellaswag_10

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hellaswag|10_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：多个路径，例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_abstract_algebra_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_anatomy_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-anatomy|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_astronomy_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-astronomy|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_business_ethics_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_clinical_knowledge_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_college_biology_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-college_biology|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_college_chemistry_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_college_computer_science_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_college_mathematics_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_college_medicine_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_college_physics_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-college_physics|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_computer_security_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-computer_security|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_conceptual_physics_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_econometrics_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-econometrics|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_electrical_engineering_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T19:19:26.317110.parquet

• harness_hendrycksTest_elementary_mathematics_5

  • 分割：2023_07_19T19_19_26.317110, latest
  • 路径：**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T19:19:26.317110.parquet

以上是数据集的概述和部分配置详情。

搜集汇总

背景与挑战

背景概述

该数据集是Open LLM Leaderboard上对模型'Fredithefish/CrimsonPajama'进行评估时自动创建的，包含64个配置对应不同任务，由2次运行生成，train分割指向最新结果，并聚合结果用于计算榜单指标，可通过datasets库加载。它用于评估和展示模型在榜单上的性能，具有结构化和自动化特点。