open-llm-leaderboard-old/details_DevaMalla__llama_7b_lora

该数据集是在评估模型DevaMalla/llama_7b_lora时自动创建的，用于Open LLM Leaderboard。它包含了64个配置，每个配置对应一个评估任务。数据集还记录了不同时间点的运行结果，并提供了一个名为results的配置，用于存储所有运行的聚合结果。此外，数据集还包括了最新的运行结果，可以通过提供的代码示例加载和查看。

原始信息汇总

数据集概述

该数据集是在对模型 DevaMalla/llama_7b_lora 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，用于计算和显示聚合指标在 Open LLM Leaderboard 上。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_DevaMalla__llama_7b_lora", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-26T14:26:37.860045 运行的最新结果：

python { "all": { "em": 0.0012583892617449664, "em_stderr": 0.0003630560893118953, "f1": 0.0611650587248323, "f1_stderr": 0.0013990352489173911, "acc": 0.3915240971461363, "acc_stderr": 0.00940445989381676 }, "harness|drop|3": { "em": 0.0012583892617449664, "em_stderr": 0.0003630560893118953, "f1": 0.0611650587248323, "f1_stderr": 0.0013990352489173911 }, "harness|gsm8k|5": { "acc": 0.05534495830174375, "acc_stderr": 0.0062982217961795855 }, "harness|winogrande|5": { "acc": 0.7277032359905288, "acc_stderr": 0.012510697991453934 } }

配置详情

• harness_arc_challenge_25

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|arc:challenge|25_2023-09-13T05-07-37.970407.parquet

• harness_drop_3

  • 分割: 2023_10_26T14_26_37.860045, latest
  • 路径: **/details_harness|drop|3_2023-10-26T14-26-37.860045.parquet

• harness_gsm8k_5

  • 分割: 2023_10_26T14_26_37.860045, latest
  • 路径: **/details_harness|gsm8k|5_2023-10-26T14-26-37.860045.parquet

• harness_hellaswag_10

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hellaswag|10_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: 多个路径，例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_abstract_algebra_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_anatomy_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_astronomy_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_business_ethics_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_clinical_knowledge_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_college_biology_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_college_chemistry_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_college_computer_science_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_college_mathematics_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_college_medicine_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_college_physics_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_computer_security_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_conceptual_physics_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_econometrics_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_electrical_engineering_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-13T05-07-37.970407.parquet

• harness_hendrycksTest_elementary_mathematics_5

  • 分割: 2023_09_13T05_07_37.970407, latest
  • 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-13T05-07-37.970407.parquet

以上是数据集的详细概述，包括数据集的组成、加载示例、最新结果以及各个配置的详细信息。