open-llm-leaderboard/details_TheBloke__wizard-mega-13B-GPTQ

该数据集是在模型TheBloke/wizard-mega-13B-GPTQ在Open LLM Leaderboard上的评估运行过程中自动创建的。数据集由3个配置组成，每个配置对应一个评估任务。数据集由2次运行创建，每次运行可以在每个配置中找到特定的分割，分割名称使用运行的时间戳。"train"分割始终指向最新的结果。额外的配置"results"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。

Dataset automatically created during the evaluation run of model TheBloke/wizard-mega-13B-GPTQ on the Open LLM Leaderboard. It is composed of 3 configurations, each corresponding to one of the evaluated tasks. The dataset includes results from 2 runs, with each run represented by a split named after the timestamp of the run. The train split always points to the latest results. An additional results configuration aggregates all the results from the runs, which are used to compute and display the aggregated metrics on the leaderboard. The dataset provides detailed metrics such as exact match (em), F1 score (f1), and accuracy (acc) for different tasks.

原始信息汇总

数据集卡片 for Evaluation run of TheBloke/wizard-mega-13B-GPTQ

数据集描述

数据集摘要

数据集是在模型 TheBloke/wizard-mega-13B-GPTQ 在 Open LLM Leaderboard 上的评估运行期间自动创建的。

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2次运行中创建。每次运行可以在每个配置中作为一个特定的分割找到，分割名称使用运行的时间戳。"train" 分割始终指向最新的结果。

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 Open LLM Leaderboard 上的聚合指标）。

加载运行细节的示例： python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_TheBloke__wizard-mega-13B-GPTQ_public", "harness_winogrande_5", split="train")

最新结果

这些是从运行 2023-11-07T07:11:46.594603 获得的最新结果：

python { "all": { "em": 0.0024119127516778523, "em_stderr": 0.0005023380498893266, "f1": 0.06481438758389294, "f1_stderr": 0.0014219270919505864, "acc": 0.4184482994878448, "acc_stderr": 0.010036365749666909 }, "harness|drop|3": { "em": 0.0024119127516778523, "em_stderr": 0.0005023380498893266, "f1": 0.06481438758389294, "f1_stderr": 0.0014219270919505864 }, "harness|gsm8k|5": { "acc": 0.08946171341925702, "acc_stderr": 0.007861583049939712 }, "harness|winogrande|5": { "acc": 0.7474348855564326, "acc_stderr": 0.012211148449394105 } }