open-llm-leaderboard/details_bigcode-data__slimpajama-1.3b

该数据集是在Open LLM Leaderboard上对bigcode-data/slimpajama-1.3b模型进行评估运行过程中自动创建的。数据集由61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，results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 generated during the evaluation run of the bigcode-data/slimpajama-1.3b model on the Open LLM Leaderboard. It comprises 61 configurations, each corresponding to a single evaluation task. The dataset is constructed from one single evaluation run, where each configuration contains a specific data split, and the name of each split is determined by the timestamp of the run. The 'train' split always points to the most recent evaluation results. Furthermore, the 'results' configuration stores the aggregated results across all runs, and is utilized to compute and display the aggregate metrics shown on the Open LLM Leaderboard.

原始信息汇总

数据集概述

数据集简介

该数据集是在对模型 bigcode-data/slimpajama-1.3b 进行评估运行期间自动创建的，用于 Open LLM Leaderboard。

数据集结构

• 配置数量：61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。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。
• 最新结果："train" 分片总是指向最新的结果。
• 汇总结果：一个额外的配置 "results" 存储所有运行的汇总结果，用于计算和显示在 Open LLM Leaderboard 上的汇总指标。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_bigcode-data__slimpajama-1.3b", "harness_truthfulqa_mc_0", split="train")

最新结果

以下是 2023-08-27T12:00:35.833850 运行 的最新结果：

python { "all": { "acc": 0.26777696601577794, "acc_stderr": 0.03181071981671739, "acc_norm": 0.2705315418772776, "acc_norm_stderr": 0.0318183562058576, "mc1": 0.23745410036719705, "mc1_stderr": 0.014896277441041834, "mc2": 0.3709121363542238, "mc2_stderr": 0.013878840992863677 }, "harness|arc:challenge|25": { "acc": 0.2781569965870307, "acc_stderr": 0.013094469919538805, "acc_norm": 0.3097269624573379, "acc_norm_stderr": 0.013512058415238361 }, "harness|hellaswag|10": { "acc": 0.42202748456482775, "acc_stderr": 0.0049287351036358335, "acc_norm": 0.5529774945230034, "acc_norm_stderr": 0.004961693567208826 }, ... }

配置详情

• harness_arc_challenge_25

  • 分片：2023_08_27T12_00_35.833850
  • 路径：**/details_harness|arc:challenge|25_2023-08-27T12:00:35.833850.parquet
  • 分片：latest
  • 路径：**/details_harness|arc:challenge|25_2023-08-27T12:00:35.833850.parquet

• harness_hellaswag_10

  • 分片：2023_08_27T12_00_35.833850
  • 路径：**/details_harness|hellaswag|10_2023-08-27T12:00:35.833850.parquet
  • 分片：latest
  • 路径：**/details_harness|hellaswag|10_2023-08-27T12:00:35.833850.parquet

• harness_hendrycksTest_5

  • 分片：2023_08_27T12_00_35.833850
  • 路径：
    • **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-27T12:00:35.833850.parquet
    • **/details_harness|hendrycksTest-anatomy|5_2023-08-27T12:00:35.833850.parquet
    • ...
  • 分片：latest
  • 路径：
    • **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-27T12:00:35.833850.parquet
    • **/details_harness|hendrycksTest-anatomy|5_2023-08-27T12:00:35.833850.parquet
    • ...