five

open-llm-leaderboard-old/details_NousResearch__Nous-Hermes-13b

收藏
Hugging Face2023-10-19 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_NousResearch__Nous-Hermes-13b
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在评估模型NousResearch/Nous-Hermes-13b时自动创建的,包含64个配置,每个配置对应一个评估任务。数据集由2次运行生成,每次运行的结果作为特定配置中的一个分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。此外,还有一个名为results的配置存储了所有运行的聚合结果,用于在Open LLM Leaderboard上计算和显示聚合指标。

该数据集是在评估模型NousResearch/Nous-Hermes-13b时自动创建的,包含64个配置,每个配置对应一个评估任务。数据集由2次运行生成,每次运行的结果作为特定配置中的一个分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。此外,还有一个名为results的配置存储了所有运行的聚合结果,用于在Open LLM Leaderboard上计算和显示聚合指标。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集简介

该数据集是在对模型 NousResearch/Nous-Hermes-13b 进行评估运行期间自动创建的,用于 Open LLM Leaderboard

数据集结构

  • 配置数量:64个配置,每个配置对应一个评估任务。
  • 运行次数:数据集从2次运行中创建。每个运行在每个配置中作为一个特定的分片存在,分片名称使用运行的时间戳。
  • 分片:每个配置包含多个分片,其中 "train" 分片始终指向最新的结果。

额外配置

  • 结果配置: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。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_NousResearch__Nous-Hermes-13b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-19T04:00:41.897332 运行的最新结果: python { "all": { "em": 0.2930998322147651, "em_stderr": 0.00466150847986569, "f1": 0.37501048657718355, "f1_stderr": 0.004576570475121802, "acc": 0.41817812997218123, "acc_stderr": 0.009868526609981134 }, "harness|drop|3": { "em": 0.2930998322147651, "em_stderr": 0.00466150847986569, "f1": 0.37501048657718355, "f1_stderr": 0.004576570475121802 }, "harness|gsm8k|5": { "acc": 0.08339651250947688, "acc_stderr": 0.00761565027710669 }, "harness|winogrande|5": { "acc": 0.7529597474348856, "acc_stderr": 0.012121402942855576 } }

配置详情

  • harness_arc_challenge_25

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-18T15:33:41.626742.parquet
  • harness_drop_3

    • 分片:2023_10_19T04_00_41.897332, latest
    • 路径:**/details_harness|drop|3_2023-10-19T04-00-41.897332.parquet
  • harness_gsm8k_5

    • 分片:2023_10_19T04_00_41.897332, latest
    • 路径:**/details_harness|gsm8k|5_2023-10-19T04-00-41.897332.parquet
  • harness_hellaswag_10

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hellaswag|10_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_anatomy_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_astronomy_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_college_biology_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_college_physics_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_computer_security_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_econometrics_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-18T15:33:41.626742.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分片:2023_07_18T15_33_41.626742, latest
    • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-18T15:33:41.626742.parquet
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