five

open-llm-leaderboard/details_mlabonne__ChimeraLlama-3-8B

收藏
Hugging Face2024-04-23 更新2024-06-12 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_mlabonne__ChimeraLlama-3-8B
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上对模型mlabonne/ChimeraLlama-3-8B进行评估时自动创建的。数据集由63个配置组成,每个配置对应一个被评估的任务。它包含一次运行的结果,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还提供了一个示例,展示了如何使用Hugging Face的datasets库加载运行中的详细信息。

该数据集是在Open LLM Leaderboard上对模型mlabonne/ChimeraLlama-3-8B进行评估时自动创建的。数据集由63个配置组成,每个配置对应一个被评估的任务。它包含一次运行的结果,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还提供了一个示例,展示了如何使用Hugging Face的datasets库加载运行中的详细信息。
提供机构:
open-llm-leaderboard
原始信息汇总

数据集概述

数据集摘要

数据集是在评估模型 mlabonne/ChimeraLlama-3-8B 的过程中自动创建的,用于 Open LLM Leaderboard。该数据集包含 63 个配置,每个配置对应一个评估任务。数据集从 1 次运行中创建,每个运行可以在每个配置中找到特定的拆分,拆分名称使用运行的时间戳。"train" 拆分始终指向最新的结果。

数据集结构

数据集包含以下配置:

  • harness_arc_challenge_25
  • harness_gsm8k_5
  • harness_hellaswag_10
  • harness_hendrycksTest_5

每个配置包含多个拆分,例如:

  • 2024_04_23T02_44_00.437694
  • latest

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_mlabonne__ChimeraLlama-3-8B", "harness_winogrande_5", split="train")

最新结果

最新结果来自运行 2024-04-23T02:44:00.437694,包含多个任务的评估结果,例如:

  • harness|arc:challenge|25
  • harness|hellaswag|10
  • harness|hendrycksTest-abstract_algebra|5
  • harness|hendrycksTest-anatomy|5
  • harness|hendrycksTest-astronomy|5
  • harness|hendrycksTest-business_ethics|5
  • harness|hendrycksTest-clinical_knowledge|5
  • harness|hendrycksTest-college_biology|5
  • harness|hendrycksTest-college_chemistry|5
  • harness|hendrycksTest-college_computer_science|5
  • harness|hendrycksTest-college_mathematics|5
  • harness|hendrycksTest-college_medicine|5
  • harness|hendrycksTest-college_physics|5
  • harness|hendrycksTest-computer_security|5
  • harness|hendrycksTest-conceptual_physics|5
  • harness|hendrycksTest-econometrics|5
  • harness|hendrycksTest-electrical_engineering|5
  • harness|hendrycksTest-elementary_mathematics|5
  • harness|hendrycksTest-formal_logic|5
  • harness|hendrycksTest-global_facts|5
  • harness|hendrycksTest-high_school_biology|5
  • harness|hendrycksTest-high_school_chemistry|5
  • harness|hendrycksTest-high_school_computer_science|5
  • harness|hendrycksTest-high_school_european_history|5
  • harness|hendrycksTest-high_school_geography|5
  • harness|hendrycksTest-high_school_government_and_politics|5
  • harness|hendrycksTest-high_school_macroeconomics|5
  • harness|hendrycksTest-high_school_mathematics|5
  • harness|hendrycksTest-high_school_microeconomics|5
  • harness|hendrycksTest-high_school_physics|5
  • harness|hendrycksTest-high_school_psychology|5
  • harness|hendrycksTest-high_school_statistics|5
  • harness|hendrycksTest-high_school_us_history|5
  • harness|hendrycksTest-high_school_world_history|5
  • harness|hendrycksTest-human_aging|5
  • harness|hendrycksTest-human_sexuality|5
  • harness|hendrycksTest-international_law|5
  • harness|hendrycksTest-jurisprudence|5
  • harness|hendrycksTest-logical_fallacies|5
  • harness|hendrycksTest-machine_learning|5
  • harness|hendrycksTest-management|5
  • harness|hendrycksTest-marketing|5
  • harness|hendrycksTest-medical_genetics|5
  • harness|hendrycksTest-miscellaneous|5
  • harness|hendrycksTest-moral_disputes|5
  • harness|hendrycksTest-moral_scenarios|5
  • harness|hendrycksTest-nutrition|5
  • harness|hendrycksTest-philosophy|5
  • harness|hendrycksTest-prehistory|5
  • harness|hendrycksTest-professional_accounting|5
  • harness|hendrycksTest-professional_law|5
  • harness|hendrycksTest-professional_medicine|5
  • harness|hendrycksTest-professional_psychology|5
  • harness|hendrycksTest-public_relations|5
  • harness|hendrycksTest-security_studies|5
  • harness|hendrycksTest-sociology|5
  • harness|hendrycksTest-us_foreign_policy|5
  • harness|hendrycksTest-virology|5
  • harness|hendrycksTest-world_religions|5
  • harness|truthfulqa:mc|0
  • harness|winogrande|5
  • harness|gsm8k|5

每个任务的结果包括准确率(acc)、标准化准确率(acc_norm)、误差(acc_stderracc_norm_stderr)等指标。

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