five

open-llm-leaderboard-old/details_nbeerbower__slerp-bophades-truthy-math-mistral-7B

收藏
Hugging Face2024-04-09 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_nbeerbower__slerp-bophades-truthy-math-mistral-7B
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nbeerbower/slerp-bophades-truthy-math-mistral-7B在Open LLM Leaderboard上的评估运行期间自动创建的。它由63个配置组成,每个配置对应一个评估任务。数据集包含1次运行的结果,每次运行在每个配置中表示为特定的分割,使用运行的时间戳命名。train分割始终指向最新结果。一个名为results的额外配置存储了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。

该数据集是在模型nbeerbower/slerp-bophades-truthy-math-mistral-7B在Open LLM Leaderboard上的评估运行期间自动创建的。它由63个配置组成,每个配置对应一个评估任务。数据集包含1次运行的结果,每次运行在每个配置中表示为特定的分割,使用运行的时间戳命名。train分割始终指向最新结果。一个名为results的额外配置存储了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。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集来源

该数据集是在对模型 nbeerbower/slerp-bophades-truthy-math-mistral-7B 进行评估运行期间自动创建的。

数据集组成

数据集由63个配置组成,每个配置对应一个评估任务。数据集从1次运行中创建,每个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。

额外配置

一个额外的配置 "results" 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。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_nbeerbower__slerp-bophades-truthy-math-mistral-7B", "harness_winogrande_5", split="train")

最新结果

以下是 最新结果 的摘要:

python { "all": { "acc": 0.6533577677120704, "acc_stderr": 0.0321090974841392, "acc_norm": 0.6524581392335448, "acc_norm_stderr": 0.032786891825831214, "mc1": 0.6242350061199511, "mc1_stderr": 0.01695458406021429, "mc2": 0.7782437262946236, "mc2_stderr": 0.0137879523668123 }, "harness|arc:challenge|25": { "acc": 0.7141638225255973, "acc_stderr": 0.013203196088537372, "acc_norm": 0.7286689419795221, "acc_norm_stderr": 0.012993807727545796 }, "harness|hellaswag|10": { "acc": 0.7197769368651663, "acc_stderr": 0.004481902637505652, "acc_norm": 0.8916550487950607, "acc_norm_stderr": 0.0031018035745563107 }, # 其他任务结果省略 }

配置详情

  • config_name: harness_arc_challenge_25

    • split: 2024_04_09T00_19_46.142948
      • path: **/details_harness|arc:challenge|25_2024-04-09T00-19-46.142948.parquet
    • split: latest
      • path: **/details_harness|arc:challenge|25_2024-04-09T00-19-46.142948.parquet
  • config_name: harness_gsm8k_5

    • split: 2024_04_09T00_19_46.142948
      • path: **/details_harness|gsm8k|5_2024-04-09T00-19-46.142948.parquet
    • split: latest
      • path: **/details_harness|gsm8k|5_2024-04-09T00-19-46.142948.parquet
  • config_name: harness_hellaswag_10

    • split: 2024_04_09T00_19_46.142948
      • path: **/details_harness|hellaswag|10_2024-04-09T00-19-46.142948.parquet
    • split: latest
      • path: **/details_harness|hellaswag|10_2024-04-09T00-19-46.142948.parquet
  • config_name: harness_hendrycksTest_5

    • split: 2024_04_09T00_19_46.142948
      • path: **/details_harness|hendrycksTest-abstract_algebra|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-anatomy|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-astronomy|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-business_ethics|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-college_biology|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-college_chemistry|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-college_computer_science|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-college_mathematics|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-college_medicine|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-college_physics|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-computer_security|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-conceptual_physics|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-econometrics|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-electrical_engineering|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-09T00-19-46.142948.parquet
      • path: **/details_harness|hendrycksTest-formal_logic|5_2024-04-09T00-19-46.142948.parquet

      其他路径省略

}

二维码
社区交流群
二维码
科研交流群
商业服务