five

open-llm-leaderboard/details_Panchovix__airoboros-33b-gpt4-1.2-SuperHOT-8k

收藏
Hugging Face2023-09-17 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_Panchovix__airoboros-33b-gpt4-1.2-SuperHOT-8k
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Panchovix/airoboros-33b-gpt4-1.2-SuperHOT-8k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,每次运行的结果作为特定配置中的一个分割,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,并用于在Open LLM Leaderboard上计算和显示聚合指标。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集来源

该数据集是在模型 Panchovix/airoboros-33b-gpt4-1.2-SuperHOT-8kOpen LLM Leaderboard 上的评估运行期间自动创建的。

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 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_Panchovix__airoboros-33b-gpt4-1.2-SuperHOT-8k", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-17T16:57:27.002060 运行的最新结果: python { "all": { "em": 0.0026216442953020135, "em_stderr": 0.0005236685642965811, "f1": 0.005930159395973156, "f1_stderr": 0.0006950327104148204, "acc": 0.2521704814522494, "acc_stderr": 0.007025978032038446 }, "harness|drop|3": { "em": 0.0026216442953020135, "em_stderr": 0.0005236685642965811, "f1": 0.005930159395973156, "f1_stderr": 0.0006950327104148204 }, "harness|gsm8k|5": { "acc": 0.0, "acc_stderr": 0.0 }, "harness|winogrande|5": { "acc": 0.5043409629044988, "acc_stderr": 0.014051956064076892 } }

配置详情

  • harness_arc_challenge_25

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|arc:challenge|25_2023-08-17T20:41:42.341199.parquet
  • harness_drop_3

    • 分片: 2023_09_17T16_57_27.002060, latest
    • 路径: **/details_harness|drop|3_2023-09-17T16-57-27.002060.parquet
  • harness_gsm8k_5

    • 分片: 2023_09_17T16_57_27.002060, latest
    • 路径: **/details_harness|gsm8k|5_2023-09-17T16-57-27.002060.parquet
  • harness_hellaswag_10

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hellaswag|10_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: 多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_anatomy_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_astronomy_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_college_biology_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_college_physics_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_computer_security_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_econometrics_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T20:41:42.341199.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分片: 2023_08_17T20_41_42.341199, latest
    • 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T20:41:42.341199.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。

二维码
社区交流群
二维码
科研交流群
商业服务