five

open-llm-leaderboard/details_MayaPH__GodziLLa-30B-plus

收藏
Hugging Face2023-08-27 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_MayaPH__GodziLLa-30B-plus
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MayaPH/GodziLLa-30B-plus进行评估运行时自动创建的。数据集由61个配置组成,每个配置对应一个评估任务。数据集是从2次运行中创建的,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。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。README还提供了如何使用Python中的datasets库加载运行细节的示例。
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 MayaPH/GodziLLa-30B-plusOpen LLM Leaderboard 上的自动创建的。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

  • 配置数量:61 个
  • 创建来源:2 次运行
  • 分割命名:每个运行的分割以时间戳命名,"train" 分割指向最新结果
  • 额外配置:"results" 存储所有运行结果的聚合,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_MayaPH__GodziLLa-30B-plus", "harness_truthfulqa_mc_0", split="train")

最新结果

最新结果来自 2023-08-02T15:59:28.566104 运行,包含多个任务的评估结果。以下是部分结果示例:

python { "all": { "acc": 0.24241616689179757, "acc_stderr": 0.031123746423736127, "acc_norm": 0.2433585961477068, "acc_norm_stderr": 0.0311383188083218, "mc1": 0.23133414932680538, "mc1_stderr": 0.014761945174862661, "mc2": 0.4855571913341874, "mc2_stderr": 0.01696461298096417 }, "harness|arc:challenge|25": { "acc": 0.23720136518771331, "acc_stderr": 0.012430399829260844, "acc_norm": 0.29180887372013653, "acc_norm_stderr": 0.013284525292403511 }, "harness|hellaswag|10": { "acc": 0.25253933479386576, "acc_stderr": 0.0043358096144803055, "acc_norm": 0.2535351523600876, "acc_norm_stderr": 0.004341454841892327 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.19, "acc_stderr": 0.03942772444036625, "acc_norm": 0.19, "acc_norm_stderr": 0.03942772444036625 }, # 其他任务结果省略... }

配置详情

  • config_name: harness_arc_challenge_25

    • splits:
      • 2023_07_28T12_09_48.036825
      • 2023_08_02T15_59_28.566104
      • latest
    • paths:
      • **/details_harness|arc:challenge|25_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|arc:challenge|25_2023-08-02T15:59:28.566104.parquet
  • config_name: harness_hellaswag_10

    • splits:
      • 2023_07_28T12_09_48.036825
      • 2023_08_02T15_59_28.566104
      • latest
    • paths:
      • **/details_harness|hellaswag|10_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hellaswag|10_2023-08-02T15:59:28.566104.parquet
  • config_name: harness_hendrycksTest_5

    • splits:
      • 2023_07_28T12_09_48.036825
      • 2023_08_02T15_59_28.566104
      • latest
    • paths:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-management|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-professional_law|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-professional_medicine|5_2023-07-28T12:09:48.036825.parquet
      • **/details_harness|hendrycksTest-professional_psychology|5_2023-07-28T12:09:48.03682
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