five

open-llm-leaderboard/details_PygmalionAI__pygmalion-2.7b

收藏
Hugging Face2023-09-22 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_PygmalionAI__pygmalion-2.7b
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PygmalionAI/pygmalion-2.7b模型进行评估时自动创建的。数据集由64个配置组成,每个配置对应一个评估任务。数据集由2次运行生成,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。此外,还有一个results配置存储了用于在排行榜上计算指标的聚合结果。可以使用`datasets`库中的`load_dataset`函数加载该数据集。

该数据集是在Open LLM Leaderboard上对PygmalionAI/pygmalion-2.7b模型进行评估时自动创建的。数据集由64个配置组成,每个配置对应一个评估任务。数据集由2次运行生成,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。此外,还有一个results配置存储了用于在排行榜上计算指标的聚合结果。可以使用`datasets`库中的`load_dataset`函数加载该数据集。
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 PygmalionAI/pygmalion-2.7b

数据集描述

该数据集是在评估模型 PygmalionAI/pygmalion-2.7bOpen LLM Leaderboard 上的自动创建的。

数据集组成

  • 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_PygmalionAI__pygmalion-2.7b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-22T20:17:59.683847 运行的最新结果: python { "all": { "em": 0.04320469798657718, "em_stderr": 0.0020821626664430564, "f1": 0.08408347315436249, "f1_stderr": 0.0023636579014392274, "acc": 0.2825572217837411, "acc_stderr": 0.006966407055209012 }, "harness|drop|3": { "em": 0.04320469798657718, "em_stderr": 0.0020821626664430564, "f1": 0.08408347315436249, "f1_stderr": 0.0023636579014392274 }, "harness|gsm8k|5": { "acc": 0.0, "acc_stderr": 0.0 }, "harness|winogrande|5": { "acc": 0.5651144435674822, "acc_stderr": 0.013932814110418024 } }

配置详情

  • harness_arc_challenge_25

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|arc:challenge|25_2023-07-19T16:36:05.422128.parquet
  • harness_drop_3

    • 分片: 2023_09_22T20_17_59.683847, latest
    • 路径: **/details_harness|drop|3_2023-09-22T20-17-59.683847.parquet
  • harness_gsm8k_5

    • 分片: 2023_09_22T20_17_59.683847, latest
    • 路径: **/details_harness|gsm8k|5_2023-09-22T20-17-59.683847.parquet
  • harness_hellaswag_10

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hellaswag|10_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: 多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_anatomy_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_astronomy_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_college_biology_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_college_physics_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_computer_security_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_econometrics_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T16:36:05.422128.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分片: 2023_07_19T16_36_05.422128, latest
    • 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T16:36:05.422128.parquet

以上是数据集的详细概述,包括数据集名称、描述、组成、加载示例、最新结果以及各个配置的详细信息。

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