open-llm-leaderboard-old/details_allenai__tulu-2-dpo-70b
收藏Hugging Face2024-02-02 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_allenai__tulu-2-dpo-70b
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在allenai/tulu-2-dpo-70b模型的评估运行过程中自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。数据集从1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。此外,还有一个名为results的配置存储了所有运行的聚合结果,并用于计算和显示在Open LLM Leaderboard上的聚合指标。
该数据集是在allenai/tulu-2-dpo-70b模型的评估运行过程中自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。数据集从1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。此外,还有一个名为results的配置存储了所有运行的聚合结果,并用于计算和显示在Open LLM Leaderboard上的聚合指标。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集名称
- 名称: Evaluation run of allenai/tulu-2-dpo-70b
数据集组成
- 配置数量: 63个配置
- 每个配置对应: 一个评估任务
- 创建来源: 从1次运行中创建
- 分割命名: 使用运行的时间戳命名
- 最新结果: "train"分割指向最新结果
额外配置
- 配置名称: "results"
- 用途: 存储所有运行的聚合结果,用于计算和显示在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_allenai__tulu-2-dpo-70b", "harness_winogrande_5", split="train")
最新结果
- 最新结果来源: 2024-02-02T06:48:43.589029
- 详细结果: 包含多个任务的准确率(acc)、标准化准确率(acc_norm)、以及其他指标(mc1, mc2等)及其标准误差(stderr)
配置详情
- 配置名称: harness_arc_challenge_25
- 数据文件路径:
**/details_harness|arc:challenge|25_2024-02-02T06-48-43.589029.parquet
- 数据文件路径:
- 配置名称: harness_gsm8k_5
- 数据文件路径:
**/details_harness|gsm8k|5_2024-02-02T06-48-43.589029.parquet
- 数据文件路径:
- 配置名称: harness_hellaswag_10
- 数据文件路径:
**/details_harness|hellaswag|10_2024-02-02T06-48-43.589029.parquet
- 数据文件路径:
- 配置名称: harness_hendrycksTest_5
- 数据文件路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-anatomy|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-astronomy|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-college_biology|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-college_physics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-computer_security|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-econometrics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-global_facts|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_geography|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_physics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-human_aging|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-human_sexuality|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-international_law|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-jurisprudence|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-machine_learning|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-management|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-marketing|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-medical_genetics|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-miscellaneous|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-moral_disputes|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-nutrition|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-philosophy|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-prehistory|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-professional_accounting|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-professional_law|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-professional_medicine|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-professional_psychology|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-public_relations|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-security_studies|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-sociology|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-us_foreign_policy|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-virology|5_2024-02-02T06-48-43.589029.parquet**/details_harness|hendrycksTest-world_religions|5_2024-02-02T06-48-43.589029.parquet
- 数据文件路径:



