open-llm-leaderboard/details_ethzanalytics__pythia-31m
收藏资源简介:
该数据集是在Open LLM Leaderboard上对模型ethzanalytics/pythia-31m进行评估时自动创建的。数据集包含64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,每次运行可以在每个配置中找到特定的分割,分割以运行的时间戳命名。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Open LLM Leaderboard上的聚合指标。
This dataset was automatically created during a single evaluation run for the model ethzanalytics/pythia-31m on the Open LLM Leaderboard. It contains 64 configurations, each corresponding to an individual evaluation task. For every configuration, specific data splits are generated within this run, with each split named after the timestamp of the run. The train split always references the most recent results. Additionally, there is a configuration named "results" that stores the aggregated results across all runs, and is used to compute and display the aggregated metrics on the Open LLM Leaderboard.
数据集概述
数据集简介
该数据集是在评估模型ethzanalytics/pythia-31m在Open LLM Leaderboard上的自动创建的。数据集包含64个配置,每个配置对应一个评估任务。
数据集结构
- 配置数量:64个配置
- 数据来源: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分割,分割名称使用运行的时间戳。
- 最新结果:"train"分割始终指向最新结果。
- 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_ethzanalytics__pythia-31m_public", "harness_winogrande_5", split="train")
最新结果
以下是2023-11-13T13:01:31.225551运行的最新结果: python { "all": { "acc": 0.2486090533214635, "acc_stderr": 0.030580280893238346, "acc_norm": 0.24951095231696532, "acc_norm_stderr": 0.031375786973211, "mc1": 0.26193390452876375, "mc1_stderr": 0.015392118805015023, "mc2": 0.49102256781530107, "mc2_stderr": 0.015750842651440947, "em": 0.0006291946308724832, "em_stderr": 0.0002568002749723811, "f1": 0.013650377516778552, "f1_stderr": 0.0006539918270891778 }, ... }
配置详情
-
harness_arc_challenge_25
- 分割:2023_11_13T13_01_31.225551, latest
- 路径:
**/details_harness|arc:challenge|25_2023-11-13T13-01-31.225551.parquet
-
harness_drop_3
- 分割:2023_11_13T13_01_31.225551, latest
- 路径:
**/details_harness|drop|3_2023-11-13T13-01-31.225551.parquet
-
harness_gsm8k_5
- 分割:2023_11_13T13_01_31.225551, latest
- 路径:
**/details_harness|gsm8k|5_2023-11-13T13-01-31.225551.parquet
-
harness_hellaswag_10
- 分割:2023_11_13T13_01_31.225551, latest
- 路径:
**/details_harness|hellaswag|10_2023-11-13T13-01-31.225551.parquet
-
harness_hendrycksTest_5
- 分割:2023_11_13T13_01_31.225551, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-11-13T13-01-31.225551.parquet, ...



