open-llm-leaderboard/details_Kiddyz__testlm-3
收藏Hugging Face2023-09-05 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard/details_Kiddyz__testlm-3
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Kiddyz/testlm-3在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由61个配置组成,每个配置对应一个被评估的任务。数据集包含1次运行的结果,每次运行作为每个配置中的特定分割存储。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,这些结果用于计算和显示Open LLM Leaderboard上的聚合指标。README还提供了一个使用Python中的datasets库加载运行细节的示例。
This dataset was automatically created during the evaluation run of the model Kiddyz/testlm-3 on the Open LLM Leaderboard. The dataset consists of 61 configurations, each corresponding to one evaluated task. The dataset contains the results of one run, 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, which are used to calculate and display the aggregate metrics on the Open LLM Leaderboard. The README also provides an example of loading run details using the datasets library in Python.
提供机构:
open-llm-leaderboard
原始信息汇总
数据集概述
数据集信息
- 名称: Evaluation run of Kiddyz/testlm-3
- 来源: 在Kiddyz/testlm-3模型在Open LLM Leaderboard上的评估运行期间自动创建。
- 组成: 包含61个配置,每个配置对应一个评估任务。
- 创建次数: 从1次运行中创建。每个运行可以在每个配置中作为一个特定的分片找到,分片名称使用运行的时间戳。"train"分片总是指向最新的结果。
- 额外配置: "results"配置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_Kiddyz__testlm-3", "harness_truthfulqa_mc_0", split="train")
最新结果
- 最新结果时间戳: 2023-09-05T01:42:44.018659
- 详细结果: 包含多个任务的准确率(acc)、标准化准确率(acc_norm)、标准误差(acc_stderr、acc_norm_stderr)等指标。
配置信息
-
配置名称: harness_arc_challenge_25
- 数据文件:
- 分片: 2023_09_05T01_42_44.018659
- 路径: **/details_harness|arc:challenge|25_2023-09-05T01:42:44.018659.parquet
- 分片: latest
- 路径: **/details_harness|arc:challenge|25_2023-09-05T01:42:44.018659.parquet
- 分片: 2023_09_05T01_42_44.018659
- 数据文件:
-
配置名称: harness_hellaswag_10
- 数据文件:
- 分片: 2023_09_05T01_42_44.018659
- 路径: **/details_harness|hellaswag|10_2023-09-05T01:42:44.018659.parquet
- 分片: latest
- 路径: **/details_harness|hellaswag|10_2023-09-05T01:42:44.018659.parquet
- 分片: 2023_09_05T01_42_44.018659
- 数据文件:
-
配置名称: harness_hendrycksTest_5
- 数据文件:
- 分片: 2023_09_05T01_42_44.018659
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-anatomy|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-astronomy|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-college_biology|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-college_physics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-computer_security|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-econometrics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-global_facts|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-05T01:42:44.018659.parquet
- **/details_harness|he
- 路径:
- 分片: 2023_09_05T01_42_44.018659
- 数据文件:



