five

open-llm-leaderboard-old/details_undi95__llama2-to-mistral-diff

收藏
Hugging Face2023-10-25 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_undi95__llama2-to-mistral-diff
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undi95/llama2-to-mistral-diff在Open LLM Leaderboard上进行评估运行期间自动创建的。数据集包含64个配置,每个配置对应一个评估任务。数据集由3次运行创建,每次运行作为一个特定分割,以运行的时间戳命名。train分割始终指向最新结果。另外,results配置存储了运行的所有聚合结果,用于在Leaderboard上计算和显示聚合指标。数据集包括针对不同任务的各种配置,每个配置具有多个针对不同时间戳和最新结果的分割。

该数据集是在模型undi95/llama2-to-mistral-diff在Open LLM Leaderboard上进行评估运行期间自动创建的。数据集包含64个配置,每个配置对应一个评估任务。数据集由3次运行创建,每次运行作为一个特定分割,以运行的时间戳命名。train分割始终指向最新结果。另外,results配置存储了运行的所有聚合结果,用于在Leaderboard上计算和显示聚合指标。数据集包括针对不同任务的各种配置,每个配置具有多个针对不同时间戳和最新结果的分割。
提供机构:
open-llm-leaderboard-old
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 undi95/llama2-to-mistral-diffOpen LLM Leaderboard 上的运行过程中自动创建的。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

  • 数据集由3次运行创建,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。
  • 额外的配置 "results" 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_undi95__llama2-to-mistral-diff", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-25T09:37:53.083823 运行的最新结果: python { "all": { "em": 0.001153523489932886, "em_stderr": 0.00034761798968571027, "f1": 0.05605494966442959, "f1_stderr": 0.0013169501309663063, "acc": 0.4076941764856182, "acc_stderr": 0.009790166925519655 }, "harness|drop|3": { "em": 0.001153523489932886, "em_stderr": 0.00034761798968571027, "f1": 0.05605494966442959, "f1_stderr": 0.0013169501309663063 }, "harness|gsm8k|5": { "acc": 0.07505686125852919, "acc_stderr": 0.007257633145486643 }, "harness|winogrande|5": { "acc": 0.7403314917127072, "acc_stderr": 0.012322700705552667 } }

配置详情

以下是数据集的配置详情:

  • harness_arc_challenge_25

    • 分割:2023_10_10T12_55_48.397880
      • 路径:**/details_harness|arc:challenge|25_2023-10-10T12-55-48.397880.parquet
    • 分割:latest
      • 路径:**/details_harness|arc:challenge|25_2023-10-10T12-55-48.397880.parquet
  • harness_drop_3

    • 分割:2023_10_24T07_59_15.869817
      • 路径:**/details_harness|drop|3_2023-10-24T07-59-15.869817.parquet
    • 分割:2023_10_25T09_37_53.083823
      • 路径:**/details_harness|drop|3_2023-10-25T09-37-53.083823.parquet
    • 分割:latest
      • 路径:**/details_harness|drop|3_2023-10-25T09-37-53.083823.parquet
  • harness_gsm8k_5

    • 分割:2023_10_24T07_59_15.869817
      • 路径:**/details_harness|gsm8k|5_2023-10-24T07-59-15.869817.parquet
    • 分割:2023_10_25T09_37_53.083823
      • 路径:**/details_harness|gsm8k|5_2023-10-25T09-37-53.083823.parquet
    • 分割:latest
      • 路径:**/details_harness|gsm8k|5_2023-10-25T09-37-53.083823.parquet
  • harness_hellaswag_10

    • 分割:2023_10_10T12_55_48.397880
      • 路径:**/details_harness|hellaswag|10_2023-10-10T12-55-48.397880.parquet
    • 分割:latest
      • 路径:**/details_harness|hellaswag|10_2023-10-10T12-55-48.397880.parquet
  • harness_hendrycksTest_5

    • 分割:2023_10_10T12_55_48.397880
      • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-formal_logic|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-global_facts|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_biology|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_geography|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_physics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-human_aging|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-human_sexuality|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-international_law|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-jurisprudence|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-machine_learning|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-management|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-marketing|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-medical_genetics|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-miscellaneous|5_2023-10-10T12-55-48.397880.parquet
      • 路径:**/details_harness|hendrycksTest-moral_disputes|5_2023-10-10T12-55-48.397880.parquet
      • 路径:`**/details_harness|hendrycksTest-moral_scenarios|5_2023
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