five

open-llm-leaderboard-old/details_bofenghuang__vigogne-2-7b-instruct

收藏
Hugging Face2023-09-23 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_bofenghuang__vigogne-2-7b-instruct
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在模型 bofenghuang/vigogne-2-7b-instruct 在 Open LLM Leaderboard 上的评估运行期间自动创建的。数据集由 64 个配置组成,每个配置对应一个被评估的任务。它包含来自 2 次运行的数据,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 上的指标。README 还提供了如何使用 `datasets` 库中的 `load_dataset` 函数加载数据集的示例。

该数据集是在模型 bofenghuang/vigogne-2-7b-instruct 在 Open LLM Leaderboard 上的评估运行期间自动创建的。数据集由 64 个配置组成,每个配置对应一个被评估的任务。它包含来自 2 次运行的数据,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 上的指标。README 还提供了如何使用 `datasets` 库中的 `load_dataset` 函数加载数据集的示例。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集来源

该数据集是在评估模型 bofenghuang/vigogne-2-7b-instructOpen LLM Leaderboard 上的运行过程中自动创建的。

数据集结构

  • 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 2 次运行中创建。每个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" 分片总是指向最新的结果。
  • 一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和显示聚合指标在 Open LLM Leaderboard 上。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_bofenghuang__vigogne-2-7b-instruct", "harness_winogrande_5", split="train")

最新结果

以下是 最新结果来自 run 2023-09-23T08:45:31.930950: python { "all": { "em": 0.2985528523489933, "em_stderr": 0.0046864904941642995, "f1": 0.3518403942953031, "f1_stderr": 0.004613402461586294, "acc": 0.39622289254314186, "acc_stderr": 0.008677803422491042 }, "harness|drop|3": { "em": 0.2985528523489933, "em_stderr": 0.0046864904941642995, "f1": 0.3518403942953031, "f1_stderr": 0.004613402461586294 }, "harness|gsm8k|5": { "acc": 0.03790750568612585, "acc_stderr": 0.005260333907798437 }, "harness|winogrande|5": { "acc": 0.7545382794001578, "acc_stderr": 0.012095272937183646 } }

配置详情

  • harness_arc_challenge_25

    • 分片: 2023_07_25T10_36_05.447803
    • 路径: **/details_harness|arc:challenge|25_2023-07-25T10:36:05.447803.parquet
    • 分片: latest
    • 路径: **/details_harness|arc:challenge|25_2023-07-25T10:36:05.447803.parquet
  • harness_drop_3

    • 分片: 2023_09_23T08_45_31.930950
    • 路径: **/details_harness|drop|3_2023-09-23T08-45-31.930950.parquet
    • 分片: latest
    • 路径: **/details_harness|drop|3_2023-09-23T08-45-31.930950.parquet
  • harness_gsm8k_5

    • 分片: 2023_09_23T08_45_31.930950
    • 路径: **/details_harness|gsm8k|5_2023-09-23T08-45-31.930950.parquet
    • 分片: latest
    • 路径: **/details_harness|gsm8k|5_2023-09-23T08-45-31.930950.parquet
  • harness_hellaswag_10

    • 分片: 2023_07_25T10_36_05.447803
    • 路径: **/details_harness|hellaswag|10_2023-07-25T10:36:05.447803.parquet
    • 分片: latest
    • 路径: **/details_harness|hellaswag|10_2023-07-25T10:36:05.447803.parquet
  • harness_hendrycksTest_5

    • 分片: 2023_07_25T10_36_05.447803
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-management|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-07-25T10:36:05.447803.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-07-25T10:36:05.447803.parquet
      • `**/details_harness|hendrycksTest-professional_law|5_2023-07-25T10:36:05.447
二维码
社区交流群
二维码
科研交流群
商业服务