open-llm-leaderboard/details_bofenghuang__vigogne-13b-instruct

该数据集是在模型bofenghuang/vigogne-13b-instruct在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由64个配置组成，每个配置对应一个评估任务。它由2次运行生成，每次运行在每个配置中表示为特定的分割。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还提供了一个如何使用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 bofenghuang/vigogne-13b-instruct on the Open LLM Leaderboard. The dataset consists of 64 configurations, each corresponding to one evaluation task. It is generated from two runs, with each run represented 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 aggregated metrics on the Open LLM Leaderboard. The README also provides an example of how to load run details using the datasets library in Python.

原始信息汇总

数据集概述

数据集摘要

该数据集是在评估模型 bofenghuang/vigogne-13b-instruct 在 Open LLM Leaderboard 上的运行过程中自动创建的。数据集包含 64 个配置，每个配置对应一个评估任务。数据集从 2 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。

数据集结构

数据集包含多个配置，每个配置对应不同的评估任务。以下是部分配置的详细信息：

配置列表

• harness_arc_challenge_25

  • 分割: 2023_07_24T14_44_50.658520, latest
  • 路径: **/details_harness|arc:challenge|25_2023-07-24T14:44:50.658520.parquet

• harness_drop_3

  • 分割: 2023_09_23T12_07_22.235368, latest
  • 路径: **/details_harness|drop|3_2023-09-23T12-07-22.235368.parquet

• harness_gsm8k_5

  • 分割: 2023_09_23T12_07_22.235368, latest
  • 路径: **/details_harness|gsm8k|5_2023-09-23T12-07-22.235368.parquet

• harness_hellaswag_10

  • 分割: 2023_07_24T14_44_50.658520, latest
  • 路径: **/details_harness|hellaswag|10_2023-07-24T14:44:50.658520.parquet

• harness_hendrycksTest_5

  • 分割: 2023_07_24T14_44_50.658520, latest
  • 路径: 多个路径，包括但不限于 **/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T14:44:50.658520.parquet

加载数据集示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_bofenghuang__vigogne-13b-instruct", "harness_winogrande_5", split="train")

最新结果

以下是 2023-09-23T12:07:22.235368 运行 的最新结果：

python { "all": { "em": 0.17365771812080538, "em_stderr": 0.003879418958892452, "f1": 0.24381711409395945, "f1_stderr": 0.003948980378514697, "acc": 0.44469214138811486, "acc_stderr": 0.010351218038230168 }, "harness|drop|3": { "em": 0.17365771812080538, "em_stderr": 0.003879418958892452, "f1": 0.24381711409395945, "f1_stderr": 0.003948980378514697 }, "harness|gsm8k|5": { "acc": 0.11827141774071266, "acc_stderr": 0.008895075852434951 }, "harness|winogrande|5": { "acc": 0.771112865035517, "acc_stderr": 0.011807360224025386 } }