five

open-llm-leaderboard/details_EleutherAI__pythia-12b

收藏
Hugging Face2023-10-12 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_EleutherAI__pythia-12b
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EleutherAI/pythia-12b模型进行评估时自动创建的。数据集由122个配置组成,每个配置对应一个被评估的任务。数据集来自4次运行,每次运行在每个配置中表示为特定的分割。train分割始终指向最新的结果。一个名为results的额外配置存储了所有运行的聚合结果,这些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`datasets`库加载数据集的示例,并包含了特定运行的最新结果。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集名称

Evaluation run of EleutherAI/pythia-12b

数据集描述

该数据集是在评估模型 EleutherAI/pythia-12bOpen LLM Leaderboard 上的自动创建的。

数据集组成

  • 数据集由 122 个配置组成,每个配置对应一个评估任务。
  • 数据集从 4 次运行中创建,每次运行可以在每个配置中作为一个特定的分片找到,分片名称使用运行的时间戳。
  • "train" 分片始终指向最新的结果。
  • 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

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_EleutherAI__pythia-12b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-12T13:49:53.203420 运行的最新结果: python { "all": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723885, "f1": 0.04447986577181216, "f1_stderr": 0.0010992181181045415, "acc": 0.32955534824940325, "acc_stderr": 0.008541034020282903 }, "harness|drop|3": { "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723885, "f1": 0.04447986577181216, "f1_stderr": 0.0010992181181045415 }, "harness|gsm8k|5": { "acc": 0.017437452615617893, "acc_stderr": 0.003605486867998272 }, "harness|winogrande|5": { "acc": 0.6416732438831886, "acc_stderr": 0.013476581172567535 } }

配置详情

  • config_name: harness_arc_challenge_25

    • split: 2023_07_19T18_06_28.460226
      • path: **/details_harness|arc:challenge|25_2023-07-19T18:06:28.460226.parquet
    • split: 2023_08_23T22_15_03.187761
      • path: **/details_harness|arc:challenge|25_2023-08-23T22:15:03.187761.parquet
    • split: latest
      • path: **/details_harness|arc:challenge|25_2023-08-23T22:15:03.187761.parquet
  • config_name: harness_drop_3

    • split: 2023_10_12T13_49_53.203420
      • path: **/details_harness|drop|3_2023-10-12T13-49-53.203420.parquet
    • split: latest
      • path: **/details_harness|drop|3_2023-10-12T13-49-53.203420.parquet
  • config_name: harness_gsm8k_5

    • split: 2023_10_12T13_49_53.203420
      • path: **/details_harness|gsm8k|5_2023-10-12T13-49-53.203420.parquet
    • split: latest
      • path: **/details_harness|gsm8k|5_2023-10-12T13-49-53.203420.parquet
  • config_name: harness_hellaswag_10

    • split: 2023_07_19T18_06_28.460226
      • path: **/details_harness|hellaswag|10_2023-07-19T18:06:28.460226.parquet
    • split: 2023_08_23T22_15_03.187761
      • path: **/details_harness|hellaswag|10_2023-08-23T22:15:03.187761.parquet
    • split: latest
      • path: **/details_harness|hellaswag|10_2023-08-23T22:15:03.187761.parquet
  • config_name: harness_hendrycksTest_5

    • split: 2023_07_19T18_06_28.460226
      • path:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-management|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-07-19T18:06:28.460226.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-07-19T18:06
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