five

open-llm-leaderboard/details_dhmeltzer__Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_merged

收藏
Hugging Face2023-10-23 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_dhmeltzer__Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_merged
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dhmeltzer/Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_merged在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由64个配置组成,每个配置对应一个被评估的任务。它包含2次运行的数据,每次运行都作为每个配置中的一个特定分割存储。train分割始终指向最新的结果。一个名为results的额外配置存储了所有运行的聚合结果,并用于计算和显示Open LLM Leaderboard上的指标。README还提供了如何使用datasets库加载数据集的示例,并包含了特定运行的最新结果。

This dataset was automatically created during the evaluation run of the model dhmeltzer/Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_merged on the Open LLM Leaderboard. The dataset consists of 64 configurations, each corresponding to one evaluated task. It contains data from two runs, with each run stored as a specific split within each configuration. The `train` split always points to the most recent results. An additional configuration named `results` stores the aggregated results across all runs, and is used to calculate and display the metrics on the Open LLM Leaderboard. The README also provides examples of how to load the dataset using the `datasets` library, and includes the latest results from specific runs.
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 dhmeltzer/Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_mergedOpen LLM Leaderboard 上的自动创建的。

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_dhmeltzer__Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_merged", "harness_winogrande_5", split="train")

最新结果

以下是 最新结果 的摘要: python { "all": { "em": 0.002307046979865772, "em_stderr": 0.0004913221265094568, "f1": 0.06630557885906033, "f1_stderr": 0.0014292163093715619, "acc": 0.4319026216914706, "acc_stderr": 0.010283324097975763 }, "harness|drop|3": { "em": 0.002307046979865772, "em_stderr": 0.0004913221265094568, "f1": 0.06630557885906033, "f1_stderr": 0.0014292163093715619 }, "harness|gsm8k|5": { "acc": 0.1068991660348749, "acc_stderr": 0.008510982565520481 }, "harness|winogrande|5": { "acc": 0.7569060773480663, "acc_stderr": 0.012055665630431043 } }

配置详情

以下是部分配置的详细信息:

  • harness_arc_challenge_25

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|arc:challenge|25_2023-10-03T17-03-59.314428.parquet
  • harness_drop_3

    • 分割: 2023_10_23T13_20_24.503429, latest
    • 路径: **/details_harness|drop|3_2023-10-23T13-20-24.503429.parquet
  • harness_gsm8k_5

    • 分割: 2023_10_23T13_20_24.503429, latest
    • 路径: **/details_harness|gsm8k|5_2023-10-23T13-20-24.503429.parquet
  • harness_hellaswag_10

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hellaswag|10_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: 多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-anatomy|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-astronomy|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-college_biology|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-college_physics|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-computer_security|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-econometrics|5_2023-10-03T17-03-59.314428.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割: 2023_10_03T17_03_59.314428, latest
    • 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-03T17-03-59.314428.parquet
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