open-llm-leaderboard/details_chargoddard__Chronorctypus-Limarobormes-13b

该数据集是在评估模型chargoddard/Chronorctypus-Limarobormes-13b时自动生成的，包含3个配置，每个配置对应一个评估任务。数据集由1次运行生成，每次运行的结果作为特定分割存储在配置中，分割名称使用运行的时间戳。train分割始终指向最新结果。此外，还有一个results配置存储所有运行的聚合结果，用于在Open LLM Leaderboard上计算和显示聚合指标。

Dataset automatically created during the evaluation run of model chargoddard/Chronorctypus-Limarobormes-13b on the Open LLM Leaderboard. The dataset is composed of 3 configurations, each corresponding to an evaluated task. The dataset has been created from 1 run(s), with each run represented as a specific split named by the timestamp. An additional results configuration stores all the aggregated results of the run and is used to compute and display the aggregated metrics on the Open LLM Leaderboard.

原始信息汇总

数据集卡片 for Evaluation run of chargoddard/Chronorctypus-Limarobormes-13b

数据集描述

数据集概述

该数据集是在模型 chargoddard/Chronorctypus-Limarobormes-13b 的评估运行期间自动创建的，用于 Open LLM Leaderboard。

数据集由3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拆分，拆分名称使用运行的时间戳。"train" 拆分始终指向最新的结果。

还有一个额外的配置 "results" 存储所有运行的聚合结果（用于计算和显示 Open LLM Leaderboard 上的聚合指标）。

最新结果

以下是来自运行 2023-10-17T10:27:33.460587 的最新结果：

python { "all": { "em": 0.05169882550335571, "em_stderr": 0.0022675304823078276, "f1": 0.17888317953020105, "f1_stderr": 0.0028882183973903902, "acc": 0.39147173871286817, "acc_stderr": 0.008785918503769254 }, "harness|drop|3": { "em": 0.05169882550335571, "em_stderr": 0.0022675304823078276, "f1": 0.17888317953020105, "f1_stderr": 0.0028882183973903902 }, "harness|gsm8k|5": { "acc": 0.03866565579984837, "acc_stderr": 0.005310583162098035 }, "harness|winogrande|5": { "acc": 0.744277821625888, "acc_stderr": 0.012261253845440473 } }

配置

• harness_drop_3

  • 拆分: 2023_10_17T10_27_33.460587
    • 路径: **/details_harness|drop|3_2023-10-17T10-27-33.460587.parquet
  • 拆分: latest
    • 路径: **/details_harness|drop|3_2023-10-17T10-27-33.460587.parquet

• harness_gsm8k_5

  • 拆分: 2023_10_17T10_27_33.460587
    • 路径: **/details_harness|gsm8k|5_2023-10-17T10-27-33.460587.parquet
  • 拆分: latest
    • 路径: **/details_harness|gsm8k|5_2023-10-17T10-27-33.460587.parquet

• harness_winogrande_5

  • 拆分: 2023_10_17T10_27_33.460587
    • 路径: **/details_harness|winogrande|5_2023-10-17T10-27-33.460587.parquet
  • 拆分: latest
    • 路径: **/details_harness|winogrande|5_2023-10-17T10-27-33.460587.parquet

• results

  • 拆分: 2023_10_17T10_27_33.460587
    • 路径: results_2023-10-17T10-27-33.460587.parquet
  • 拆分: latest
    • 路径: results_2023-10-17T10-27-33.460587.parquet