five

open-llm-leaderboard/details_wenge-research__yayi-7b-llama2

收藏
Hugging Face2023-09-23 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_wenge-research__yayi-7b-llama2
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wenge-research/yayi-7b-llama2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次运行创建,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
原始信息汇总

数据集概述

数据集摘要

该数据集是在评估模型 wenge-research/yayi-7b-llama2Open LLM Leaderboard 上的自动创建的。数据集包含 64 个配置,每个配置对应一个评估任务。

数据集结构

  • 配置数量:64 个配置
  • 配置类型:每个配置对应一个评估任务
  • 数据来源:数据集从 4 次运行中创建,每次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

  • 结果配置:名为 "results" 的额外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

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_wenge-research__yayi-7b-llama2", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-23T15:31:48.459973 运行的最新结果: python { "all": { "em": 0.0017827181208053692, "em_stderr": 0.0004320097346038788, "f1": 0.06044777684563769, "f1_stderr": 0.0013803417345618757, "acc": 0.40589214880805274, "acc_stderr": 0.009561073756915776 }, "harness|drop|3": { "em": 0.0017827181208053692, "em_stderr": 0.0004320097346038788, "f1": 0.06044777684563769, "f1_stderr": 0.0013803417345618757 }, "harness|gsm8k|5": { "acc": 0.0667172100075815, "acc_stderr": 0.006873340544455132 }, "harness|winogrande|5": { "acc": 0.745067087608524, "acc_stderr": 0.012248806969376422 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_07_24T11_30_18.986210
      • 路径:**/details_harness|arc:challenge|25_2023-07-24T11:30:18.986210.parquet
    • 分割:2023_07_27T10_53_28.461116
      • 路径:**/details_harness|arc:challenge|25_2023-07-27T10:53:28.461116.parquet
    • 分割:latest
      • 路径:**/details_harness|arc:challenge|25_2023-07-27T10:53:28.461116.parquet
  • harness_drop_3

    • 分割:2023_09_23T09_26_06.826895
      • 路径:**/details_harness|drop|3_2023-09-23T09-26-06.826895.parquet
    • 分割:2023_09_23T15_31_48.459973
      • 路径:**/details_harness|drop|3_2023-09-23T15-31-48.459973.parquet
    • 分割:latest
      • 路径:**/details_harness|drop|3_2023-09-23T15-31-48.459973.parquet
  • harness_gsm8k_5

    • 分割:2023_09_23T09_26_06.826895
      • 路径:**/details_harness|gsm8k|5_2023-09-23T09-26-06.826895.parquet
    • 分割:2023_09_23T15_31_48.459973
      • 路径:**/details_harness|gsm8k|5_2023-09-23T15-31-48.459973.parquet
    • 分割:latest
      • 路径:**/details_harness|gsm8k|5_2023-09-23T15-31-48.459973.parquet
  • harness_hellaswag_10

    • 分割:2023_07_24T11_30_18.986210
      • 路径:**/details_harness|hellaswag|10_2023-07-24T11:30:18.986210.parquet
    • 分割:2023_07_27T10_53_28.461116
      • 路径:**/details_harness|hellaswag|10_2023-07-27T10:53:28.461116.parquet
    • 分割:latest
      • 路径:**/details_harness|hellaswag|10_2023-07-27T10:53:28.461116.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_24T11_30_18.986210
      • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-formal_logic|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-global_facts|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_biology|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_european_history|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_geography|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_physics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_psychology|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_statistics|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_us_history|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_world_history|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-human_aging|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-human_sexuality|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-international_law|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-jurisprudence|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-logical_fallacies|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-machine_learning|5_2023-07-24T11:30:18.986210.parquet
      • 路径:**/details_harness|hendrycksTest-management|5_2023-07-24T11:30:18.986210.parquet
      • 路径:`**/details_harness|hendrycksTest-marketing|5_2023-07-24T11:30:18.9862
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