open-llm-leaderboard-old/details_ddobokki__Llama-2-70b-orca-200k

该数据集是在评估模型ddobokki/Llama-2-70b-orca-200k时自动创建的，评估在Open LLM Leaderboard上进行。数据集由61个配置组成，每个配置对应一个评估任务。数据集是从1次运行中创建的，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。

原始信息汇总

数据集概述

数据集信息

• 数据集名称: Evaluation run of ddobokki/Llama-2-70b-orca-200k
• 数据集来源: 自动创建于模型 ddobokki/Llama-2-70b-orca-200k 在 Open LLM Leaderboard 的评估运行。
• 数据集组成: 包含61个配置，每个配置对应一个评估任务。
• 数据集创建: 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。"train" 分割始终指向最新结果。
• 额外配置: "results" 配置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。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_ddobokki__Llama-2-70b-orca-200k", "harness_truthfulqa_mc_0", split="train")

最新结果

• 最新结果时间戳: 2023-08-09T20:21:28.711089
• 聚合结果: python { "all": { "acc": 0.6675525003199225, "acc_stderr": 0.0320256356518761, "acc_norm": 0.6716808545277895, "acc_norm_stderr": 0.03199912887877205, "mc1": 0.408812729498164, "mc1_stderr": 0.01720995215164173, "mc2": 0.5618014117500216, "mc2_stderr": 0.015000194909320638 }, "harness|arc:challenge|25": { "acc": 0.5989761092150171, "acc_stderr": 0.014322255790719867, "acc_norm": 0.6484641638225256, "acc_norm_stderr": 0.013952413699600935 }, "harness|hellaswag|10": { "acc": 0.6584345747858992, "acc_stderr": 0.004732654295724444, "acc_norm": 0.8525194184425413, "acc_norm_stderr": 0.0035385967737048313 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.38, "acc_stderr": 0.048783173121456316, "acc_norm": 0.38, "acc_norm_stderr": 0.048783173121456316 }, "harness|hendrycksTest-anatomy|5": { "acc": 0.6148148148148148, "acc_stderr": 0.042039210401562783, "acc_norm": 0.6148148148148148, "acc_norm_stderr": 0.042039210401562783 }, "harness|hendrycksTest-astronomy|5": { "acc": 0.7697368421052632, "acc_stderr": 0.03426059424403165, "acc_norm": 0.7697368421052632, "acc_norm_stderr": 0.03426059424403165 }, "harness|hendrycksTest-business_ethics|5": { "acc": 0.71, "acc_stderr": 0.04560480215720684, "acc_norm": 0.71, "acc_norm_stderr": 0.04560480215720684 }, "harness|hendrycksTest-clinical_knowledge|5": { "acc": 0.6981132075471698, "acc_stderr": 0.02825420034443866, "acc_norm": 0.6981132075471698, "acc_norm_stderr": 0.02825420034443866 }, "harness|hendrycksTest-college_biology|5": { "acc": 0.7847222222222222, "acc_stderr": 0.034370793441061344, "acc_norm": 0.7847222222222222, "acc_norm_stderr": 0.034370793441061344 }, "harness|hendrycksTest-college_chemistry|5": { "acc": 0.45, "acc_stderr": 0.04999999999999999, "acc_norm": 0.45, "acc_norm_stderr": 0.04999999999999999 }, "harness|hendrycksTest-college_computer_science|5": { "acc": 0.57, "acc_stderr": 0.04975698519562428, "acc_norm": 0.57, "acc_norm_stderr": 0.04975698519562428 }, "harness|hendrycksTest-college_mathematics|5": { "acc": 0.36, "acc_stderr": 0.04824181513244218, "acc_norm": 0.36, "acc_norm_stderr": 0.04824181513244218 }, "harness|hendrycksTest-college_medicine|5": { "acc": 0.6127167630057804, "acc_stderr": 0.037143259063020656, "acc_norm": 0.6127167630057804, "acc_norm_stderr": 0.037143259063020656 }, "harness|hendrycksTest-college_physics|5": { "acc": 0.3431372549019608, "acc_stderr": 0.04724007352383888, "acc_norm": 0.3431372549019608, "acc_norm_stderr": 0.04724007352383888 }, "harness|hendrycksTest-computer_security|5": { "acc": 0.74, "acc_stderr": 0.04408440022768078, "acc_norm": 0.74, "acc_norm_stderr": 0.04408440022768078 }, "harness|hendrycksTest-conceptual_physics|5": { "acc": 0.6127659574468085, "acc_stderr": 0.03184389265339526, "acc_norm": 0.6127659574468085, "acc_norm_stderr": 0.03184389265339526 }, "harness|hendrycksTest-econometrics|5": { "acc": 0.4298245614035088, "acc_stderr": 0.04657047260594962, "acc_norm": 0.4298245614035088, "acc_norm_stderr": 0.04657047260594962 }, "harness|hendrycksTest-electrical_engineering|5": { "acc": 0.593103448275862, "acc_stderr": 0.04093793981266236, "acc_norm": 0.593103448275862, "acc_norm_stderr": 0.04093793981266236 }, "harness|hendrycksTest-elementary_mathematics|5": { "acc": 0.41534391534391535, "acc_stderr": 0.025379524910778408, "acc_norm": 0.41534391534391535, "acc_norm_stderr": 0.025379524910778408 }, "harness|hendrycksTest-formal_logic|5": { "acc": 0.4365079365079365, "acc_stderr": 0.04435932892851466, "acc_norm": 0.4365079365079365, "acc_norm_stderr": 0.04435932892851466 }, "harness|hendrycksTest-global_facts|5": { "acc": 0.43, "acc_stderr": 0.049756985195624284, "acc_norm": 0.43, "acc_norm_stderr": 0.049756985195624284 }, "harness|hendrycksTest-high_school_biology|5": { "acc": 0.8129032258064516, "acc_stderr": 0.02218571009225225, "acc_norm": 0.8129032258064516, "acc_norm_stderr": 0.02218571009225225 }, "harness|hendrycksTest-high_school_chemistry|5": { "acc": 0.5123152709359606, "acc_stderr": 0.035169204442208966, "acc_norm": 0.5123152709359606, "acc_norm_stderr": 0.035169204442208966 }, "harness|hendrycksTest-high_school_computer_science|5": { "acc": 0.72, "acc_stderr": 0.04512608598542127, "acc_norm": 0.72, "acc_norm_stderr": 0.04512608598542127 }, "harness|hendrycksTest-high_school_european_history|5": { "acc": 0.8121212121212121, "acc_stderr": 0.03050193405942914, "acc_norm": 0.8121212121212121, "acc_norm_stderr": 0.03050193405942914 }, "harness|hendrycksTest-high_school_geography|5": { "acc": 0.8636363636363636, "acc_stderr": 0.024450155973189835, "acc_norm": 0.8636363636363636, "acc_norm_stderr": 0.024450155973189835 }, "harness|hendrycksTest-high_school_government_and_politics|5": { "acc": 0.9067357512953368, "acc_stderr": 0.0209868545932897, "acc_norm": 0.9067357512953368, "acc_norm_stderr": 0.0209868545932897 }, "harness|hendrycksTest-high_school_macroeconomics|5": { "acc": 0.6820512820512821, "acc_stderr": 0.02361088430892786, "acc_norm": 0.6820512820512821, "acc_norm_stderr": 0.02361088430892786 }, "harness|hendrycksTest-high_school_mathematics|5": { "acc": 0.35555555555555557, "acc_stderr": 0.02918571494985741, "acc_norm": 0.35555555555555557, "acc_norm_stderr": 0.02918571494985741

搜集汇总

背景与挑战

背景概述

该数据集是评估模型'ddobokki/Llama-2-70b-orca-200k'时自动生成的，包含61个配置对应不同评估任务，用于在Open LLM Leaderboard上计算和显示聚合指标。它基于单次运行，通过时间戳分割结果，train分割始终指向最新评估数据。