five

open-llm-leaderboard/details_elyza__ELYZA-japanese-Llama-2-7b

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_elyza__ELYZA-japanese-Llama-2-7b
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在评估模型elyza/ELYZA-japanese-Llama-2-7b时自动创建的,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,每次运行的结果存储在不同的分割中,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,存储所有运行的聚合结果,用于在Open LLM Leaderboard上计算和显示聚合指标。

This dataset was automatically created during the evaluation of the model elyza/ELYZA-japanese-Llama-2-7b. It includes 64 configurations, each corresponding to one evaluation task. Generated from two runs, the results of each run are stored in separate dataset splits, where the split names use the timestamp of the corresponding run. The "train" split always points to the most recent results. Additionally, there is a configuration named "results" that stores the aggregated results across all runs, which are used to compute and display aggregate metrics on the Open LLM Leaderboard.
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

该数据集是在评估模型 elyza/ELYZA-japanese-Llama-2-7bOpen LLM Leaderboard 上的运行过程中自动创建的。数据集包含 64 个配置,每个配置对应一个评估任务。

数据集结构

数据集由 2 次运行创建,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" 存储了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。

加载数据示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_elyza__ELYZA-japanese-Llama-2-7b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-17T01:32:48.448739 运行的最新结果

python { "all": { "em": 0.0009437919463087249, "em_stderr": 0.00031446531194133837, "f1": 0.05525901845637585, "f1_stderr": 0.0013597649565375795, "acc": 0.4051552412586848, "acc_stderr": 0.010068835191488079 }, "harness|drop|3": { "em": 0.0009437919463087249, "em_stderr": 0.00031446531194133837, "f1": 0.05525901845637585, "f1_stderr": 0.0013597649565375795 }, "harness|gsm8k|5": { "acc": 0.08339651250947688, "acc_stderr": 0.007615650277106702 }, "harness|winogrande|5": { "acc": 0.7269139700078927, "acc_stderr": 0.012522020105869456 } }

配置详情

以下是数据集的配置详情:

  • config_name: harness_arc_challenge_25

    • split: 2023_08_31T10_22_13.739402
      • path: **/details_harness|arc:challenge|25_2023-08-31T10:22:13.739402.parquet
    • split: latest
      • path: **/details_harness|arc:challenge|25_2023-08-31T10:22:13.739402.parquet
  • config_name: harness_drop_3

    • split: 2023_09_17T01_32_48.448739
      • path: **/details_harness|drop|3_2023-09-17T01-32-48.448739.parquet
    • split: latest
      • path: **/details_harness|drop|3_2023-09-17T01-32-48.448739.parquet
  • config_name: harness_gsm8k_5

    • split: 2023_09_17T01_32_48.448739
      • path: **/details_harness|gsm8k|5_2023-09-17T01-32-48.448739.parquet
    • split: latest
      • path: **/details_harness|gsm8k|5_2023-09-17T01-32-48.448739.parquet
  • config_name: harness_hellaswag_10

    • split: 2023_08_31T10_22_13.739402
      • path: **/details_harness|hellaswag|10_2023-08-31T10:22:13.739402.parquet
    • split: latest
      • path: **/details_harness|hellaswag|10_2023-08-31T10:22:13.739402.parquet
  • config_name: harness_hendrycksTest_5

    • split: 2023_08_31T10_22_13.739402
      • path:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-management|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-08-31T10:22:13.739402.parquet
        • **/details_harness|hendrycksTest-professional_law|5_2023-08-31T10:22:1
搜集汇总
数据集介绍
main_image_url
以上内容由遇见数据集搜集并总结生成
二维码
社区交流群
二维码
科研交流群
商业服务