open-llm-leaderboard/details_prince-canuma__Damysus-Coder-v0.1

该数据集是在Open LLM Leaderboard上对模型prince-canuma/Damysus-Coder-v0.1进行评估时自动创建的。数据集包含63个配置，每个配置对应一个评估任务。数据集由1次运行创建，每次运行的结果作为一个特定的分割存储在配置中，分割名称使用运行的时间戳。train分割始终指向最新的结果。此外，还有一个名为results的配置存储了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。

原始信息汇总

数据集概述

数据集名称: Evaluation run of prince-canuma/Damysus-Coder-v0.1

数据集描述: 该数据集是在评估模型prince-canuma/Damysus-Coder-v0.1运行期间自动创建的，用于Open LLM Leaderboard。

数据集组成:

• 配置数量: 63个
• 每个配置对应任务: 每个配置对应一个评估任务
• 数据集来源: 来自1次运行
• 数据集分割: 每个运行作为特定分割，分割名称使用运行的时间戳命名
• 特殊配置: “results”配置存储所有运行结果的聚合结果，用于计算和显示聚合指标

数据集加载示例: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_prince-canuma__Damysus-Coder-v0.1", "harness_winogrande_5", split="train")

最新结果

最新结果来自2024-04-15T21:07:57.637768的运行，包含多个任务的评估结果，例如：

• harness|arc:challenge|25: 准确率(acc)为0.5554607508532423
• harness|hellaswag|10: 准确率(acc)为0.64070902210715
• harness|hendrycksTest-abstract_algebra|5: 准确率(acc)为0.3
• harness|hendrycksTest-anatomy|5: 准确率(acc)为0.5703703703703704
• harness|hendrycksTest-astronomy|5: 准确率(acc)为0.6118421052631579
• harness|hendrycksTest-business_ethics|5: 准确率(acc)为0.57
• harness|hendrycksTest-clinical_knowledge|5: 准确率(acc)为0.690566037735849
• harness|hendrycksTest-college_biology|5: 准确率(acc)为0.6527777777777778
• harness|hendrycksTest-college_chemistry|5: 准确率(acc)为0.37
• harness|hendrycksTest-college_computer_science|5: 准确率(acc)为0.49
• harness|hendrycksTest-college_mathematics|5: 准确率(acc)为0.36
• harness|hendrycksTest-college_medicine|5: 准确率(acc)为0.5780346820809249
• harness|hendrycksTest-college_physics|5: 准确率(acc)为0.46078431372549017
• harness|hendrycksTest-computer_security|5: 准确率(acc)为0.74
• harness|hendrycksTest-conceptual_physics|5: 准确率(acc)为0.5191489361702127
• harness|hendrycksTest-econometrics|5: 准确率(acc)为0.40350877192982454
• harness|hendrycksTest-electrical_engineering|5: 准确率(acc)为0.5793103448275863
• harness|hendrycksTest-elementary_mathematics|5: 准确率(acc)为0.3835978835978836
• harness|hendrycksTest-formal_logic|5: 准确率(acc)为0.36507936507936506
• harness|hendrycksTest-global_facts|5: 准确率(acc)为0.4
• harness|hendrycksTest-high_school_biology|5: 准确率(acc)为0.6967741935483871
• harness|hendrycksTest-high_school_chemistry|5: 准确率(acc)为0.5073891625615764
• harness|hendrycksTest-high_school_computer_science|5: 准确率(acc)为0.63
• harness|hendrycksTest-high_school_european_history|5: 准确率(acc)为0.7333333333333333
• harness|hendrycksTest-high_school_geography|5: 准确率(acc)为0.7424242424242424
• harness|hendrycksTest-high_school_government_and_politics|5: 准确率(acc)为0.8549222797927462
• harness|hendrycksTest-high_school_macroeconomics|5: 准确率(acc)为0.5615384615384615
• harness|hendrycksTest-high_school_mathematics|5: 准确率(acc)为0.34444444444444444
• harness|hendrycksTest-high_school_microeconomics|5: 准确率(acc)为0.6302521008403361
• harness|hendrycksTest-high_school_physics|5: 准确率(acc)为0.3509933774834437
• harness|hendrycksTest-high_school_psychology|5: 准确率(acc)为0.8
• harness|hendrycksTest-high_school_statistics|5: 准确率(acc)为0.4444444444444444
• harness|hendrycksTest-high_school_us_history|5: 准确率(acc)为0.7745098039215687
• harness|hendrycksTest-high_school_world_history|5: 准确率(acc)为0.7510548523206751
• harness|hendrycksTest-human_aging|5: 准确率(acc)为0.6367713004484304
• harness|hendrycksTest-human_sexuality|5: 准确率(acc)为0.7022900763358778
• harness|hendrycksTest-international_law|5: 准确率(acc)为0.8099173553719008
• harness|hendrycksTest-jurisprudence|5: 准确率(acc)为0.6851851851851852
• harness|hendrycksTest-logical_fallacies|5: 准确率(acc)为0.7300613496932515
• harness|hendrycksTest-machine_learning|5: 准确率(acc)为0.48214285714285715
• harness|hendrycksTest-management|5: 准确率(acc)为0.7475728155339806
• harness|hendrycksTest-marketing|5: 准确率(acc)为0.8717948717948718
• harness|hendrycksTest-medical_genetics|5: 准确率(acc)为0.68
• harness|hendrycksTest-miscellaneous|5: 准确率(acc)为0.7739463601532567
• harness|hendrycksTest-moral_disputes|5: 准确率(acc)为0.6791907514450867
• harness|hendrycksTest-moral_scenarios|5: 准确率(acc)为0.3396648044692737
• harness|hendrycksTest-nutrition|5: 准确率(acc)为0.7058823529411765
• harness|hendrycksTest-philosophy|5: 准确率(acc)为0.6881028938906752
• harness|hendrycksTest-prehistory|5: 准确率(acc)为0.6697530864197531
• harness|hendrycksTest-professional_accounting|5: 准确率(acc)为0.4574468085106383
• harness|hendrycksTest-professional_law|5: 准确率(acc)为0.42633637548891784
• harness|hendrycksTest-professional_medicine|5: 准确率(acc)为0.6029411764705882
• harness|hendrycksTest-professional_psychology|5: 准确率(acc)为0.6143790849673203
• harness|hendrycksTest-public_relations|5: 准确率(acc)为0.7
• harness|hendrycksTest-security_studies|5: 准确率(acc)为0.7387755102040816
• harness|hendrycksTest-sociology|5: 准确率(acc)为0.746268656716418
• harness|hendrycksTest-us_foreign_policy|5: 准确率(acc)为0.81
• harness|hendrycksTest-virology|5: 准确率(acc)为0.5180722891566265
• harness|hendrycksTest-world_religions|5: 准确率(acc)为0.8245614035087719
• harness|truthfulqa:mc|0: mc1为0.4724602203182375，mc2为0.6419919334749323
• harness|winogrande|5: 准确率(acc)为0.771112865035517
• harness|gsm8k|5: 准确率(acc)为0.39272175890826383

以上数据集概述提供了模型评估结果的详细信息，包括各个任务的准确率等指标。