five

open-llm-leaderboard/details_KoboldAI__PPO_Pygway-6b-Mix

收藏
Hugging Face2023-10-19 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_KoboldAI__PPO_Pygway-6b-Mix
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在评估模型KoboldAI/PPO_Pygway-6b-Mix时自动创建的,包含64个配置,每个配置对应一个评估任务。数据集由2次运行生成,每次运行的结果存储在不同的分割中,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,数据集还包含一个名为results的配置,用于存储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。

This dataset was automatically created during the evaluation of the model KoboldAI/PPO_Pygway-6b-Mix, containing 64 configurations, each corresponding to one evaluation task. The dataset is generated from two runs, with results of each run stored in separate splits named after the run's timestamp. The `train` split always points to the results of the most recent run. Additionally, the dataset includes a configuration named `results` that stores the aggregated results across all runs, and is used to calculate and display the aggregated metrics on the Open LLM Leaderboard.
提供机构:
open-llm-leaderboard
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 KoboldAI/PPO_Pygway-6b-MixOpen LLM Leaderboard 上的自动创建的。

数据集结构

  • 数据集由 64 个配置组成,每个配置对应一个评估任务。
  • 数据集从 2 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_KoboldAI__PPO_Pygway-6b-Mix", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-19T08:04:51.574728 运行的最新结果: python { "all": { "em": 0.0010486577181208054, "em_stderr": 0.0003314581465219131, "f1": 0.05169567953020148, "f1_stderr": 0.0012499480042451514, "acc": 0.33036017216649627, "acc_stderr": 0.008492168272498208 }, "harness|drop|3": { "em": 0.0010486577181208054, "em_stderr": 0.0003314581465219131, "f1": 0.05169567953020148, "f1_stderr": 0.0012499480042451514 }, "harness|gsm8k|5": { "acc": 0.016679302501895376, "acc_stderr": 0.0035275958887224603 }, "harness|winogrande|5": { "acc": 0.6440410418310971, "acc_stderr": 0.013456740656273954 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T15:47:31.801752.parquet
  • harness_drop_3

    • 分割:2023_10_19T08_04_51.574728, latest
    • 路径:**/details_harness|drop|3_2023-10-19T08-04-51.574728.parquet
  • harness_gsm8k_5

    • 分割:2023_10_19T08_04_51.574728, latest
    • 路径:**/details_harness|gsm8k|5_2023-10-19T08-04-51.574728.parquet
  • harness_hellaswag_10

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hellaswag|10_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T15:47:31.801752.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分割:2023_07_19T15_47_31.801752, latest
    • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T15:47:31.801752.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