five

open-llm-leaderboard/details_psmathur__orca_mini_v3_7b

收藏
Hugging Face2023-10-18 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_psmathur__orca_mini_v3_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上对模型psmathur/orca_mini_v3_7b进行评估时自动生成的。数据集包含64个配置,每个配置对应一个被评估的任务。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分割,分割名称由运行的时间戳命名。train分割始终指向最新的结果。此外,一个名为results的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上的聚合指标。README还提供了如何使用Hugging Face datasets库加载运行中的详细信息的示例。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集来源

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,每次运行在每个配置中作为一个特定的分片存在,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" 存储所有运行结果的聚合,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_psmathur__orca_mini_v3_7b", "harness_winogrande_5", split="train")

最新结果

  • 最新结果来自 2023-10-18T04:27:15.231240 运行,包含以下指标: python { "all": { "em": 0.08043204697986577, "em_stderr": 0.0027851341980506704, "f1": 0.15059563758389252, "f1_stderr": 0.0030534563383277672, "acc": 0.4069827001752661, "acc_stderr": 0.009686225873410097 }, "harness|drop|3": { "em": 0.08043204697986577, "em_stderr": 0.0027851341980506704, "f1": 0.15059563758389252, "f1_stderr": 0.0030534563383277672 }, "harness|gsm8k|5": { "acc": 0.0712661106899166, "acc_stderr": 0.007086462127954491 }, "harness|winogrande|5": { "acc": 0.7426992896606156, "acc_stderr": 0.012285989618865706 } }

配置详情

  • harness_arc_challenge_25

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|arc:challenge|25_2023-08-16T13:35:32.670682.parquet
  • harness_drop_3

    • 分片: 2023_10_18T04_27_15.231240, latest
    • 路径: **/details_harness|drop|3_2023-10-18T04-27-15.231240.parquet
  • harness_gsm8k_5

    • 分片: 2023_10_18T04_27_15.231240, latest
    • 路径: **/details_harness|gsm8k|5_2023-10-18T04-27-15.231240.parquet
  • harness_hellaswag_10

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hellaswag|10_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: 多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_anatomy_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_astronomy_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_college_biology_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_college_physics_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_computer_security_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_econometrics_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-16T13:35:32.670682.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分片: 2023_08_16T13_35_32.670682, latest
    • 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-16T13:35:32.670682.parquet

以上是数据集的详细概述,包括数据集的来源、结构、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