five

open-llm-leaderboard/details_huggingface__llama-13b

收藏
Hugging Face2023-09-23 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_huggingface__llama-13b
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在Open LLM Leaderboard上对huggingface/llama-13b模型进行评估时自动生成的。数据集包含64个配置,每个配置对应一个评估任务。数据集由2次运行生成,每次运行的结果作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,并用于在Open LLM Leaderboard上计算和显示聚合指标。

This dataset was automatically generated during the evaluation of the huggingface/llama-13b model on the Open LLM Leaderboard. It comprises 64 configurations, each corresponding to one evaluation task. The dataset is generated from two runs, with each run's results acting as a split under its respective configuration, where the split name is the timestamp of the corresponding run. The train split always points to the results of the most recent run. Additionally, there is a configuration named "results" that stores the aggregated results of all runs, and is used to calculate 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

数据集名称

  • Evaluation run of huggingface/llama-13b

数据集描述

  • 该数据集是在对模型 huggingface/llama-13b 进行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。
  • 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_huggingface__llama-13b", "harness_winogrande_5", split="train")

最新结果

  • 这些是最新的结果,来自 2023-09-23T00:11:57.766802 的运行: python { "all": { "em": 0.0019924496644295304, "em_stderr": 0.000456667646266702, "f1": 0.056602348993288636, "f1_stderr": 0.0013004668300984712, "acc": 0.4191229752993855, "acc_stderr": 0.009626252314482865 }, "harness|drop|3": { "em": 0.0019924496644295304, "em_stderr": 0.000456667646266702, "f1": 0.056602348993288636, "f1_stderr": 0.0013004668300984712 }, "harness|gsm8k|5": { "acc": 0.0758150113722517, "acc_stderr": 0.007291205723162579 }, "harness|winogrande|5": { "acc": 0.7624309392265194, "acc_stderr": 0.011961298905803152 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-20T10:00:13.254246.parquet
  • harness_drop_3

    • 分割:2023_09_23T00_11_57.766802, latest
    • 路径:**/details_harness|drop|3_2023-09-23T00-11-57.766802.parquet
  • harness_gsm8k_5

    • 分割:2023_09_23T00_11_57.766802, latest
    • 路径:**/details_harness|gsm8k|5_2023-09-23T00-11-57.766802.parquet
  • harness_hellaswag_10

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hellaswag|10_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-20T10:00:13.254246.parquet
  • harness_hendrycksTest_elementary_mathematics_5

    • 分割:2023_07_20T10_00_13.254246, latest
    • 路径:**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-20T10:00:13.254246.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