five

open-llm-leaderboard/details_CHIH-HUNG__llama-2-13b-FINETUNE5_4w-r8-gate_up_down

收藏
Hugging Face2023-10-28 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_CHIH-HUNG__llama-2-13b-FINETUNE5_4w-r8-gate_up_down
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CHIH-HUNG/llama-2-13b-FINETUNE5_4w-r8-gate_up_down的评估运行期间自动创建的,用于Open LLM Leaderboard的评估。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2次运行生成,每次运行的结果作为特定分割存储,分割名称使用运行的时间戳。train分割始终指向最新结果。此外,还有一个名为results的配置,存储了所有运行的聚合结果,用于计算和显示Open LLM Leaderboard上的聚合指标。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集来源

该数据集是在评估模型 CHIH-HUNG/llama-2-13b-FINETUNE5_4w-r8-gate_up_downOpen LLM Leaderboard 上的运行过程中自动创建的。

数据集结构

  • 该数据集包含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。

额外配置

  • 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 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_CHIH-HUNG__llama-2-13b-FINETUNE5_4w-r8-gate_up_down", "harness_winogrande_5", split="train")

最新结果

这些是最新结果(来自2023-10-28T04:49:22.682759): python { "all": { "em": 0.07487416107382551, "em_stderr": 0.0026952933607895307, "f1": 0.12380872483221442, "f1_stderr": 0.00285148758396334, "acc": 0.44832731261215925, "acc_stderr": 0.010657041987495935 }, "harness|drop|3": { "em": 0.07487416107382551, "em_stderr": 0.0026952933607895307, "f1": 0.12380872483221442, "f1_stderr": 0.00285148758396334 }, "harness|gsm8k|5": { "acc": 0.133434420015163, "acc_stderr": 0.00936649160978448 }, "harness|winogrande|5": { "acc": 0.7632202052091555, "acc_stderr": 0.01194759236520739 } }

配置详情

  • harness_arc_challenge_25

    • 分割: 2023_10_10T11_25_01.199069
    • 路径: **/details_harness|arc:challenge|25_2023-10-10T11-25-01.199069.parquet
    • 分割: latest
    • 路径: **/details_harness|arc:challenge|25_2023-10-10T11-25-01.199069.parquet
  • harness_drop_3

    • 分割: 2023_10_28T04_49_22.682759
    • 路径: **/details_harness|drop|3_2023-10-28T04-49-22.682759.parquet
    • 分割: latest
    • 路径: **/details_harness|drop|3_2023-10-28T04-49-22.682759.parquet
  • harness_gsm8k_5

    • 分割: 2023_10_28T04_49_22.682759
    • 路径: **/details_harness|gsm8k|5_2023-10-28T04-49-22.682759.parquet
    • 分割: latest
    • 路径: **/details_harness|gsm8k|5_2023-10-28T04-49-22.682759.parquet
  • harness_hellaswag_10

    • 分割: 2023_10_10T11_25_01.199069
    • 路径: **/details_harness|hellaswag|10_2023-10-10T11-25-01.199069.parquet
    • 分割: latest
    • 路径: **/details_harness|hellaswag|10_2023-10-10T11-25-01.199069.parquet
  • harness_hendrycksTest_5

    • 分割: 2023_10_10T11_25_01.199069
    • 路径:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_biology|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_european_history|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_geography|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_physics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_psychology|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_statistics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_us_history|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-high_school_world_history|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-human_aging|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-human_sexuality|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-international_law|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-jurisprudence|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-logical_fallacies|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-machine_learning|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-management|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-marketing|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-medical_genetics|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-miscellaneous|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-moral_disputes|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-moral_scenarios|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-nutrition|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-philosophy|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-prehistory|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-professional_accounting|5_2023-10-10T11-25-01.199069.parquet
      • **/details_harness|hendrycksTest-professional_law|5_2023-10-10T11-25-01.199069.parquet
      • `**/details_harness|hendrycksTest-professional_medicine|5_2023-10-1
二维码
社区交流群
二维码
科研交流群
商业服务