five

open-llm-leaderboard-old/details_AA051610__T2A

收藏
Hugging Face2023-10-11 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_AA051610__T2A
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在Open LLM Leaderboard上对模型AA051610/T2A进行评估时自动创建的。数据集由61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拆分,train拆分始终指向最新结果。此外,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Open LLM Leaderboard上的聚合指标。文件还提供了如何使用`datasets`库中的`load_dataset`函数加载运行详情的示例。

该数据集是在Open LLM Leaderboard上对模型AA051610/T2A进行评估时自动创建的。数据集由61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拆分,train拆分始终指向最新结果。此外,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Open LLM Leaderboard上的聚合指标。文件还提供了如何使用`datasets`库中的`load_dataset`函数加载运行详情的示例。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集简介

该数据集是在模型AA051610/T2AOpen LLM Leaderboard上的评估运行期间自动创建的。数据集包含61个配置,每个配置对应一个评估任务。

数据集结构

数据集由以下部分组成:

  • 配置:每个配置对应一个评估任务。
  • 分割:每个配置包含多个分割,分割名称使用运行的时间戳。"train"分割始终指向最新结果。
  • 结果配置:名为"results"的配置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

以下是加载数据集详细信息的示例代码: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_AA051610__T2A", "harness_truthfulqa_mc_0", split="train")

最新结果

以下是最新结果的摘要: python { "all": { "acc": 0.6173946376864352, "acc_stderr": 0.03337871209335492, "acc_norm": 0.6210628259045844, "acc_norm_stderr": 0.03336880945880684, "mc1": 0.31334149326805383, "mc1_stderr": 0.0162380650690596, "mc2": 0.47014420938426915, "mc2_stderr": 0.014571966148559557 }, "harness|arc:challenge|25": { "acc": 0.4854948805460751, "acc_stderr": 0.014605241081370053, "acc_norm": 0.514505119453925, "acc_norm_stderr": 0.014605241081370056 }, "harness|hellaswag|10": { "acc": 0.5524795857398924, "acc_stderr": 0.0049622205125483525, "acc_norm": 0.739892451702848, "acc_norm_stderr": 0.004377965074211627 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.22, "acc_stderr": 0.041633319989322695, "acc_norm": 0.22, "acc_norm_stderr": 0.041633319989322695 }, # 其他任务的结果... }

配置详情

以下是部分配置的详细信息:

  • harness_arc_challenge_25

    • 分割:2023_10_11T15_16_23.487044
    • 路径:**/details_harness|arc:challenge|25_2023-10-11T15-16-23.487044.parquet
    • 分割:latest
    • 路径:**/details_harness|arc:challenge|25_2023-10-11T15-16-23.487044.parquet
  • harness_hellaswag_10

    • 分割:2023_10_11T15_16_23.487044
    • 路径:**/details_harness|hellaswag|10_2023-10-11T15-16-23.487044.parquet
    • 分割:latest
    • 路径:**/details_harness|hellaswag|10_2023-10-11T15-16-23.487044.parquet
  • harness_hendrycksTest_5

    • 分割:2023_10_11T15_16_23.487044
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-management|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-professional_law|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-professional_medicine|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-professional_psychology|5_2023-10-11T15-16-23.487044.parquet
      • **/details_harness|hendrycksTest-public_relations|5_2023-10-11T15-16-23.487044.parquet
      • `**/details_harness|hendrycksTest-security
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作