five

open-llm-leaderboard/details_lgaalves__tinyllama-1.1b-chat-v0.3_platypus

收藏
Hugging Face2023-10-23 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_lgaalves__tinyllama-1.1b-chat-v0.3_platypus
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在评估模型[lgaalves/tinyllama-1.1b-chat-v0.3_platypus]在[Open LLM Leaderboard]上的表现时自动创建的。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一个特定的分割存储,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,一个名为results的配置存储了所有运行的聚合结果,用于计算和显示在[Open LLM Leaderboard]上的聚合指标。

该数据集是在评估模型[lgaalves/tinyllama-1.1b-chat-v0.3_platypus]在[Open LLM Leaderboard]上的表现时自动创建的。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一个特定的分割存储,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,一个名为results的配置存储了所有运行的聚合结果,用于计算和显示在[Open LLM Leaderboard]上的聚合指标。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

该数据集是在评估模型lgaalves/tinyllama-1.1b-chat-v0.3_platypusOpen LLM Leaderboard上的运行过程中自动创建的。

数据集组成

  • 该数据集包含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"分割始终指向最新的结果。
  • 一个额外的配置"results"存储所有运行结果的聚合,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_lgaalves__tinyllama-1.1b-chat-v0.3_platypus", "harness_winogrande_5", split="train")

最新结果

以下是最新结果来自2023-10-23T23:05:04.270048: python { "all": { "em": 0.0025167785234899327, "em_stderr": 0.0005131152834514911, "f1": 0.049414848993288615, "f1_stderr": 0.0012773102707031435, "acc": 0.2816590502599073, "acc_stderr": 0.00797944490002852 }, "harness|drop|3": { "em": 0.0025167785234899327, "em_stderr": 0.0005131152834514911, "f1": 0.049414848993288615, "f1_stderr": 0.0012773102707031435 }, "harness|gsm8k|5": { "acc": 0.00530705079605762, "acc_stderr": 0.002001305720948044 }, "harness|winogrande|5": { "acc": 0.5580110497237569, "acc_stderr": 0.013957584079108994 } }

配置详情

  • harness_arc_challenge_25

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|arc:challenge|25_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|arc:challenge|25_2023-10-10T14-53-56.428911.parquet
  • harness_drop_3

    • 分割: 2023_10_23T23_05_04.270048
    • 路径: **/details_harness|drop|3_2023-10-23T23-05-04.270048.parquet
    • 分割: latest
    • 路径: **/details_harness|drop|3_2023-10-23T23-05-04.270048.parquet
  • harness_gsm8k_5

    • 分割: 2023_10_23T23_05_04.270048
    • 路径: **/details_harness|gsm8k|5_2023-10-23T23-05-04.270048.parquet
    • 分割: latest
    • 路径: **/details_harness|gsm8k|5_2023-10-23T23-05-04.270048.parquet
  • harness_hellaswag_10

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hellaswag|10_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hellaswag|10_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T14-53-56.428911.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-10-10T14-53-56.428911.parquet
      • ...(其他路径省略)
    • 分割: latest
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T14-53-56.428911.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-10-10T14-53-56.428911.parquet
      • ...(其他路径省略)
  • harness_hendrycksTest_abstract_algebra_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-10T14-53-56.428911.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割: 2023_10_10T14_53_56.428911
    • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-10-10T14-53-56.428911.parquet
    • 分割: latest
    • 路径: `**/details_harness
二维码
社区交流群
二维码
科研交流群
商业服务