five

open-llm-leaderboard/details_upstage__Llama-2-70b-instruct

收藏
Hugging Face2023-10-17 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_upstage__Llama-2-70b-instruct
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upstage/Llama-2-70b-instruct的评估运行期间自动创建的,用于Open LLM Leaderboard的评估。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,每次运行的结果存储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,用于计算和显示Open LLM Leaderboard上的聚合指标。

This dataset was automatically generated during the evaluation runs of the model upstage/Llama-2-70b-instruct, and is intended for evaluations on the Open LLM Leaderboard. The dataset consists of 64 configurations, each corresponding to a single evaluation task. It is derived from two separate runs, where the results of each run are stored as a split under a specific configuration, with the split name being the timestamp of the corresponding run. The "train" split always references the most up-to-date results. Additionally, there is a configuration named "results" that stores the aggregated results across all runs, which are used to compute and display the aggregated metrics on the Open LLM Leaderboard.
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集简介

该数据集是在对模型 upstage/Llama-2-70b-instruct 进行评估运行期间自动创建的,用于 Open LLM Leaderboard

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。
  • 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_upstage__Llama-2-70b-instruct", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-17T12:48:24.237609 运行的最新结果: python { "all": { "em": 0.49989513422818793, "em_stderr": 0.005120467878578845, "f1": 0.5841736577181234, "f1_stderr": 0.004671177225967014, "acc": 0.5754715400500128, "acc_stderr": 0.011730426388075654 }, "harness|drop|3": { "em": 0.49989513422818793, "em_stderr": 0.005120467878578845, "f1": 0.5841736577181234, "f1_stderr": 0.004671177225967014 }, "harness|gsm8k|5": { "acc": 0.32221379833206976, "acc_stderr": 0.01287243548118878 }, "harness|winogrande|5": { "acc": 0.8287292817679558, "acc_stderr": 0.010588417294962526 } }

配置详情

以下是数据集的配置详情:

  • harness_arc_challenge_25

    • 分割:2023_07_31T16_38_35.808290
      • 路径:**/details_harness|arc:challenge|25_2023-07-31T16:38:35.808290.parquet
    • 分割:latest
      • 路径:**/details_harness|arc:challenge|25_2023-07-31T16:38:35.808290.parquet
  • harness_drop_3

    • 分割:2023_10_17T12_48_24.237609
      • 路径:**/details_harness|drop|3_2023-10-17T12-48-24.237609.parquet
    • 分割:latest
      • 路径:**/details_harness|drop|3_2023-10-17T12-48-24.237609.parquet
  • harness_gsm8k_5

    • 分割:2023_10_17T12_48_24.237609
      • 路径:**/details_harness|gsm8k|5_2023-10-17T12-48-24.237609.parquet
    • 分割:latest
      • 路径:**/details_harness|gsm8k|5_2023-10-17T12-48-24.237609.parquet
  • harness_hellaswag_10

    • 分割:2023_07_31T16_38_35.808290
      • 路径:**/details_harness|hellaswag|10_2023-07-31T16:38:35.808290.parquet
    • 分割:latest
      • 路径:**/details_harness|hellaswag|10_2023-07-31T16:38:35.808290.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_31T16_38_35.808290
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-management|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-07-31T16:38:35.808290.parquet
        • **/details_harness|hendrycksTest-professional_law|5_2023-07-31T16:38:35.808290.parquet
        • `**/details_harness|hendrycksTest-professional
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