five

open-llm-leaderboard-old/details_duoqi__Nanbeige-16B-Base-Llama

收藏
Hugging Face2024-01-18 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_duoqi__Nanbeige-16B-Base-Llama
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duoqi/Nanbeige-16B-Base-Llama的评估运行中自动创建的,包含63个配置,每个配置对应一个评估任务。数据集是从1次运行中创建的,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。此外,还有一个名为results的配置,存储了所有运行的聚合结果,并用于计算和显示在Open LLM Leaderboard上的聚合指标。

该数据集是在模型duoqi/Nanbeige-16B-Base-Llama的评估运行中自动创建的,包含63个配置,每个配置对应一个评估任务。数据集是从1次运行中创建的,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。此外,还有一个名为results的配置,存储了所有运行的聚合结果,并用于计算和显示在Open LLM Leaderboard上的聚合指标。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集简介

该数据集是在评估模型duoqi/Nanbeige-16B-Base-LlamaOpen LLM Leaderboard上的自动创建的。数据集包含63个配置,每个配置对应一个评估任务。

数据集结构

数据集由1次运行创建,每个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"分割总是指向最新的结果。

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。

加载数据示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_duoqi__Nanbeige-16B-Base-Llama", "harness_winogrande_5", split="train")

最新结果

以下是最新结果(来自2024-01-18T13:45:09.426165的运行): python { "all": { "acc": 0.6301701399529482, "acc_stderr": 0.032386700810762455, "acc_norm": 0.6349032587315631, "acc_norm_stderr": 0.03303120442847387, "mc1": 0.29865361077111385, "mc1_stderr": 0.016021570613768542, "mc2": 0.42602438122791003, "mc2_stderr": 0.014551892220412064 }, "harness|arc:challenge|25": { "acc": 0.5213310580204779, "acc_stderr": 0.014598087973127106, "acc_norm": 0.5648464163822525, "acc_norm_stderr": 0.014487986197186045 }, "harness|hellaswag|10": { "acc": 0.5865365465046803, "acc_stderr": 0.004914480534533712, "acc_norm": 0.7896833300139414, "acc_norm_stderr": 0.004067006345542834 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.31, "acc_stderr": 0.04648231987117316, "acc_norm": 0.31, "acc_norm_stderr": 0.04648231987117316 }, ... }

配置详情

  • config_name: harness_arc_challenge_25

    • data_files:
      • split: 2024_01_18T13_45_09.426165
        • path:
          • **/details_harness|arc:challenge|25_2024-01-18T13-45-09.426165.parquet
      • split: latest
        • path:
          • **/details_harness|arc:challenge|25_2024-01-18T13-45-09.426165.parquet
  • config_name: harness_gsm8k_5

    • data_files:
      • split: 2024_01_18T13_45_09.426165
        • path:
          • **/details_harness|gsm8k|5_2024-01-18T13-45-09.426165.parquet
      • split: latest
        • path:
          • **/details_harness|gsm8k|5_2024-01-18T13-45-09.426165.parquet
  • config_name: harness_hellaswag_10

    • data_files:
      • split: 2024_01_18T13_45_09.426165
        • path:
          • **/details_harness|hellaswag|10_2024-01-18T13-45-09.426165.parquet
      • split: latest
        • path:
          • **/details_harness|hellaswag|10_2024-01-18T13-45-09.426165.parquet
  • config_name: harness_hendrycksTest_5

    • data_files:
      • split: 2024_01_18T13_45_09.426165
        • path:
          • **/details_harness|hendrycksTest-abstract_algebra|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-anatomy|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-astronomy|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-business_ethics|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-college_biology|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-college_chemistry|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-college_computer_science|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-college_mathematics|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-college_medicine|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-college_physics|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-computer_security|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-conceptual_physics|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-econometrics|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-electrical_engineering|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-formal_logic|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-global_facts|5_2024-01-18T13-45-09.426165.parquet
          • **/details_harness|hendrycksTest-high_school_biology|5_2024-01-18T13-45-09.426165.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