five

open-llm-leaderboard-old/details_CHIH-HUNG__llama-2-13b-FINETUNE4_3.8w-r4-gate_up_down

收藏
Hugging Face2023-10-25 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_CHIH-HUNG__llama-2-13b-FINETUNE4_3.8w-r4-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-FINETUNE4_3.8w-r4-gate_up_down](https://huggingface.co/CHIH-HUNG/llama-2-13b-FINETUNE4_3.8w-r4-gate_up_down)在[Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)上评估运行期间自动创建的。数据集包含64个配置,每个配置对应一个评估任务。每个配置中包含使用运行时间戳命名的分割,其中train分割始终指向最新结果。此外,results配置存储了运行的聚合结果,用于计算和显示leaderboard上的聚合指标。数据集结构设计便于加载特定运行细节,如提供的Python代码片段所示。

该数据集是在模型[CHIH-HUNG/llama-2-13b-FINETUNE4_3.8w-r4-gate_up_down](https://huggingface.co/CHIH-HUNG/llama-2-13b-FINETUNE4_3.8w-r4-gate_up_down)在[Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)上评估运行期间自动创建的。数据集包含64个配置,每个配置对应一个评估任务。每个配置中包含使用运行时间戳命名的分割,其中train分割始终指向最新结果。此外,results配置存储了运行的聚合结果,用于计算和显示leaderboard上的聚合指标。数据集结构设计便于加载特定运行细节,如提供的Python代码片段所示。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

该数据集是在评估模型 CHIH-HUNG/llama-2-13b-FINETUNE4_3.8w-r4-gate_up_downOpen LLM Leaderboard 上的运行过程中自动创建的。

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。每次运行可以在每个配置中作为一个特定的分割找到,分割名称使用运行的时间戳。
  • "train" 分割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_CHIH-HUNG__llama-2-13b-FINETUNE4_3.8w-r4-gate_up_down", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-25T22:05:19.521411 运行的最新结果

python { "all": { "em": 0.2774748322147651, "em_stderr": 0.004585407529667309, "f1": 0.3228513003355709, "f1_stderr": 0.004521161118733854, "acc": 0.4456737872141304, "acc_stderr": 0.010576995457593712 }, "harness|drop|3": { "em": 0.2774748322147651, "em_stderr": 0.004585407529667309, "f1": 0.3228513003355709, "f1_stderr": 0.004521161118733854 }, "harness|gsm8k|5": { "acc": 0.1281273692191054, "acc_stderr": 0.009206398549980031 }, "harness|winogrande|5": { "acc": 0.7632202052091555, "acc_stderr": 0.011947592365207394 } }

配置详情

  • harness_arc_challenge_25

    • 分割: 2023_10_01T14_49_36.997866
      • 路径: **/details_harness|arc:challenge|25_2023-10-01T14-49-36.997866.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2023-10-01T14-49-36.997866.parquet
  • harness_drop_3

    • 分割: 2023_10_25T22_05_19.521411
      • 路径: **/details_harness|drop|3_2023-10-25T22-05-19.521411.parquet
    • 分割: latest
      • 路径: **/details_harness|drop|3_2023-10-25T22-05-19.521411.parquet
  • harness_gsm8k_5

    • 分割: 2023_10_25T22_05_19.521411
      • 路径: **/details_harness|gsm8k|5_2023-10-25T22-05-19.521411.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2023-10-25T22-05-19.521411.parquet
  • harness_hellaswag_10

    • 分割: 2023_10_01T14_49_36.997866
      • 路径: **/details_harness|hellaswag|10_2023-10-01T14-49-36.997866.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2023-10-01T14-49-36.997866.parquet
  • harness_hendrycksTest_5

    • 分割: 2023_10_01T14_49_36.997866
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-management|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-10-01T14-49-36.997866.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-10-01T14-49-36.997866.parquet
        • `**/details_harness|hendrycksTest-professional_accounting|5_2023-10-01T14-49-36.9
二维码
社区交流群
二维码
科研交流群
商业服务