open-llm-leaderboard-old/details_1TuanPham__T-Llama
收藏Hugging Face2024-03-29 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_1TuanPham__T-Llama
下载链接
链接失效反馈官方服务:
资源简介:
数据集是在评估模型1TuanPham/T-Llama时自动创建的,评估运行在Open LLM Leaderboard上进行。数据集由63个配置组成,每个配置对应一个评估任务。数据集是从1次运行中创建的,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,用于计算和显示Open LLM Leaderboard上的聚合指标。
数据集是在评估模型1TuanPham/T-Llama时自动创建的,评估运行在Open LLM Leaderboard上进行。数据集由63个配置组成,每个配置对应一个评估任务。数据集是从1次运行中创建的,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,用于计算和显示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
数据集组成
- 该数据集包含63个配置,每个配置对应一个评估任务。
- 数据集由1次运行创建,每个运行的详细结果存储在特定的分割中,分割名称使用运行的时间戳。
- 每个配置中的“train”分割指向最新的结果。
- 额外的配置“results”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_1TuanPham__T-Llama", "harness_winogrande_5", split="train")
最新结果
- 最新结果来自2024-03-29T21:13:24.001614的运行,包含多个任务的准确率和标准误差。
- 示例结果: python { "all": { "acc": 0.48260613994162815, "acc_stderr": 0.034597958283579845, "acc_norm": 0.4856702378089888, "acc_norm_stderr": 0.03532936344005532, "mc1": 0.3072215422276622, "mc1_stderr": 0.016150201321323006, "mc2": 0.464708469527708, "mc2_stderr": 0.015628743915417735 }, "harness|arc:challenge|25": { "acc": 0.5034129692832765, "acc_stderr": 0.014611050403244077, "acc_norm": 0.5418088737201365, "acc_norm_stderr": 0.0145602203087147 }, "harness|hellaswag|10": { "acc": 0.5784704242182832, "acc_stderr": 0.004927948061486069, "acc_norm": 0.7647878908583947, "acc_norm_stderr": 0.004232645108976139 }, # 其他任务的结果... }
配置详情
- 数据集包含多个配置,每个配置对应不同的任务和数据文件路径。
- 示例配置:
config_name: harness_arc_challenge_25split: 2024_03_29T21_13_24.001614path: **/details_harness|arc:challenge|25_2024-03-29T21-13-24.001614.parquet
split: latestpath: **/details_harness|arc:challenge|25_2024-03-29T21-13-24.001614.parquet
config_name: harness_gsm8k_5split: 2024_03_29T21_13_24.001614path: **/details_harness|gsm8k|5_2024-03-29T21-13-24.001614.parquet
split: latestpath: **/details_harness|gsm8k|5_2024-03-29T21-13-24.001614.parquet
config_name: harness_hellaswag_10split: 2024_03_29T21_13_24.001614path: **/details_harness|hellaswag|10_2024-03-29T21-13-24.001614.parquet
split: latestpath: **/details_harness|hellaswag|10_2024-03-29T21-13-24.001614.parquet
config_name: harness_hendrycksTest_5split: 2024_03_29T21_13_24.001614path: **/details_harness|hendrycksTest-abstract_algebra|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-anatomy|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-astronomy|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-business_ethics|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-clinical_knowledge|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-college_biology|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-college_chemistry|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-college_computer_science|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-college_mathematics|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-college_medicine|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-college_physics|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-computer_security|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-conceptual_physics|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-econometrics|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-electrical_engineering|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-elementary_mathematics|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-formal_logic|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-global_facts|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-high_school_biology|5_2024-03-29T21-13-24.001614.parquetpath: **/details_harness|hendrycksTest-high_school_chemistry|5_2024-03-29T21-13-24.001614.parquet
其他任务的路径...



