five

open-llm-leaderboard/details_facebook__opt-13b

收藏
Hugging Face2023-12-02 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_facebook__opt-13b
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facebook/opt-13b模型进行评估时自动创建的。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4次运行生成,每次运行的结果作为特定分割存储在配置中,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,还有一个名为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

数据集创建背景

该数据集是在对模型 facebook/opt-13b 进行评估运行期间自动创建的,评估结果展示在 Open LLM Leaderboard 上。

数据集组成

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 4 次运行中创建,每次运行在每个配置中作为一个特定的分割存在,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

以下是加载特定运行详细信息的示例代码: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_facebook__opt-13b", "harness_gsm8k_5", split="train")

最新结果

以下是 2023-12-02T16:45:23.193106 运行的最新结果: python { "all": { "acc": 0.017437452615617893, "acc_stderr": 0.0036054868679982863 }, "harness|gsm8k|5": { "acc": 0.017437452615617893, "acc_stderr": 0.0036054868679982863 } }

配置详情

配置列表

  • harness_arc_challenge_25

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T18:53:15.812068.parquet
  • harness_drop_3

    • 分割:2023_09_09T11_29_00.543425, 2023_10_13T09_11_13.715169, latest
    • 路径:**/details_harness|drop|3_2023-09-09T11-29-00.543425.parquet, **/details_harness|drop|3_2023-10-13T09-11-13.715169.parquet
  • harness_gsm8k_5

    • 分割:2023_09_09T11_29_00.543425, 2023_10_13T09_11_13.715169, 2023_12_02T16_45_23.193106, latest
    • 路径:**/details_harness|gsm8k|5_2023-09-09T11-29-00.543425.parquet, **/details_harness|gsm8k|5_2023-10-13T09-11-13.715169.parquet, **/details_harness|gsm8k|5_2023-12-02T16-45-23.193106.parquet
  • harness_hellaswag_10

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hellaswag|10_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:多个路径,详见原文
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_formal_logic_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_global_facts_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_biology_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_chemistry_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_computer_science_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_european_history_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_geography_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_government_and_politics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_macroeconomics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_mathematics_5

    • 分割:2023_07_19T18_53_15.812068, latest
    • 路径:**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:53:15.812068.parquet
  • harness_hendrycksTest_high_school_microeconomics_5

    • 分割:2023_07_1
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