five

open-llm-leaderboard-old/details_kevin009__babyllama-v0.6

收藏
Hugging Face2024-02-13 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_kevin009__babyllama-v0.6
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在模型kevin009/babyllama-v0.6在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。它包含一次运行的结果,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。README还提供了如何使用Hugging Face datasets库加载数据集的示例。

该数据集是在模型kevin009/babyllama-v0.6在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。它包含一次运行的结果,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。README还提供了如何使用Hugging Face datasets库加载数据集的示例。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集摘要

该数据集是在对模型 kevin009/babyllama-v0.6 进行评估运行期间自动创建的,用于 Open LLM Leaderboard

数据集组成

  • 该数据集包含 63 个配置,每个配置对应一个评估任务。
  • 数据集从 1 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_kevin009__babyllama-v0.6", "harness_winogrande_5", split="train")

最新结果

以下是 2024-02-13T10:06:30.565512 运行的最新结果

python { "all": { "acc": 0.26066347798834766, "acc_stderr": 0.030904794820091792, "acc_norm": 0.26161932329960463, "acc_norm_stderr": 0.0316608342460649, "mc1": 0.22031823745410037, "mc1_stderr": 0.014509045171487295, "mc2": 0.3584100057903431, "mc2_stderr": 0.013776314892170112 }, "harness|arc:challenge|25": { "acc": 0.35238907849829354, "acc_stderr": 0.013960142600598677, "acc_norm": 0.3609215017064846, "acc_norm_stderr": 0.014034761386175458 }, "harness|hellaswag|10": { "acc": 0.46335391356303524, "acc_stderr": 0.004976361454341339, "acc_norm": 0.6159131647082254, "acc_norm_stderr": 0.0048538457503921415 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.24, "acc_stderr": 0.04292346959909284, "acc_norm": 0.24, "acc_norm_stderr": 0.04292346959909284 }, ... }

配置详情

  • config_name: harness_arc_challenge_25

    • split: 2024_02_13T10_06_30.565512
      • path: **/details_harness|arc:challenge|25_2024-02-13T10-06-30.565512.parquet
    • split: latest
      • path: **/details_harness|arc:challenge|25_2024-02-13T10-06-30.565512.parquet
  • config_name: harness_gsm8k_5

    • split: 2024_02_13T10_06_30.565512
      • path: **/details_harness|gsm8k|5_2024-02-13T10-06-30.565512.parquet
    • split: latest
      • path: **/details_harness|gsm8k|5_2024-02-13T10-06-30.565512.parquet
  • config_name: harness_hellaswag_10

    • split: 2024_02_13T10_06_30.565512
      • path: **/details_harness|hellaswag|10_2024-02-13T10-06-30.565512.parquet
    • split: latest
      • path: **/details_harness|hellaswag|10_2024-02-13T10-06-30.565512.parquet
  • config_name: harness_hendrycksTest_5

    • split: 2024_02_13T10_06_30.565512
      • path:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2024-02-13T10-06-30.565512.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2024-02-13T10-06-30.565512.parquet ...
二维码
社区交流群
二维码
科研交流群
商业服务