five

open-llm-leaderboard/details_budecosystem__boomer-1b

收藏
Hugging Face2023-10-24 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_budecosystem__boomer-1b
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 budecosystem/boomer-1b 在 Open LLM Leaderboard 上进行评估时自动创建的。数据集由 64 个配置组成,每个配置对应一个被评估的任务。数据集是从 2 次运行中生成的,每次运行在每个配置中表示为特定的分割。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,这些结果用于计算和显示 Open LLM Leaderboard 上的聚合指标。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 generated during the evaluation of the model budecosystem/boomer-1b on the Open LLM Leaderboard. It consists of 64 configurations, each corresponding to a single evaluation task. The dataset is derived from two experimental runs, where each run is represented as a dedicated split under each configuration. The "train" split always references the most up-to-date evaluation results. An additional configuration named "results" stores the aggregated results across all runs, which are utilized to compute and display the aggregate metrics on the Open LLM Leaderboard. The README also includes examples of loading the dataset via the `datasets` library, alongside the most recent results from specific experimental runs.
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集来源

该数据集是在对模型 budecosystem/boomer-1b 进行评估运行期间自动创建的,评估结果展示在 Open LLM Leaderboard 上。

数据集结构

  • 配置数量:64个配置,每个配置对应一个评估任务。
  • 创建次数:数据集从2次运行中创建。每个运行在每个配置中作为一个特定的分割存在,分割名称使用运行的时间戳。
  • 最新结果:"train" 分割总是指向最新的结果。
  • 聚合结果:一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和展示在 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_budecosystem__boomer-1b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-24T14:53:25.007106 运行 的最新结果: python { "all": { "em": 0.0007340604026845638, "em_stderr": 0.0002773614457335763, "f1": 0.052141359060402785, "f1_stderr": 0.0013172260484977333, "acc": 0.2571140151259026, "acc_stderr": 0.008333536236283095 }, "harness|drop|3": { "em": 0.0007340604026845638, "em_stderr": 0.0002773614457335763, "f1": 0.052141359060402785, "f1_stderr": 0.0013172260484977333 }, "harness|gsm8k|5": { "acc": 0.009097801364670205, "acc_stderr": 0.002615326510775672 }, "harness|winogrande|5": { "acc": 0.505130228887135, "acc_stderr": 0.014051745961790516 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|arc:challenge|25_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|arc:challenge|25_2023-10-09T15-37-37.200624.parquet
  • harness_drop_3

    • 分割:2023_10_24T14_53_25.007106
    • 路径:**/details_harness|drop|3_2023-10-24T14-53-25.007106.parquet
    • 分割:latest
    • 路径:**/details_harness|drop|3_2023-10-24T14-53-25.007106.parquet
  • harness_gsm8k_5

    • 分割:2023_10_24T14_53_25.007106
    • 路径:**/details_harness|gsm8k|5_2023-10-24T14-53-25.007106.parquet
    • 分割:latest
    • 路径:**/details_harness|gsm8k|5_2023-10-24T14-53-25.007106.parquet
  • harness_hellaswag_10

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hellaswag|10_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hellaswag|10_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:多个路径,例如 **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-10-09T15-37-37.200624.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_10_09T15_37_37.200624
    • 路径:**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-09T15-37-37.200624.parquet
    • 分割:latest
    • 路径
搜集汇总
背景与挑战
背景概述
该数据集是模型budecosystem/boomer-1b在Open LLM Leaderboard评估时自动生成的评估结果集合,包含64个配置对应不同任务,从2次运行中产生,train分割提供最新结果,results配置用于计算排行榜聚合指标,便于用户通过datasets库加载和分析模型性能。
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