open-llm-leaderboard-old/details_HenryJJ__Instruct_Yi-6B_Dolly_CodeAlpaca
收藏数据集概述
数据集摘要
该数据集是在评估模型 HenryJJ/Instruct_Yi-6B_Dolly_CodeAlpaca 在 Open LLM Leaderboard 上的自动创建的。数据集包含 63 个配置,每个配置对应一个评估任务。
数据集结构
- 创建来源: 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。
- 额外配置: 一个名为 "results" 的额外配置存储所有运行的聚合结果,用于计算和显示在 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_HenryJJ__Instruct_Yi-6B_Dolly_CodeAlpaca", "harness_winogrande_5", split="train")
最新结果
以下是 2024-01-08T03:36:26.320528 运行的最新结果:
python { "all": { "acc": 0.6236571632946254, "acc_stderr": 0.03222013618820489, "acc_norm": 0.6309524776297984, "acc_norm_stderr": 0.03288521739348617, "mc1": 0.27906976744186046, "mc1_stderr": 0.0157021070906279, "mc2": 0.41422968964840373, "mc2_stderr": 0.014212709995879808 }, "harness|arc:challenge|25": { "acc": 0.5025597269624573, "acc_stderr": 0.014611199329843784, "acc_norm": 0.5315699658703071, "acc_norm_stderr": 0.014582236460866975 }, "harness|hellaswag|10": { "acc": 0.5616411073491336, "acc_stderr": 0.004951717622007979, "acc_norm": 0.7530372435769767, "acc_norm_stderr": 0.0043036354511158045 }, # 其他任务结果省略 }
配置详情
-
配置名称: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_01_07T21_59_12.253105
- 路径: **/details_harness|arc:challenge|25_2024-01-07T21-59-12.253105.parquet
- 分割: 2024_01_08T03_36_26.320528
- 路径: **/details_harness|arc:challenge|25_2024-01-08T03-36-26.320528.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2024-01-08T03-36-26.320528.parquet
- 分割: 2024_01_07T21_59_12.253105
- 数据文件:
-
配置名称: harness_gsm8k_5
- 数据文件:
- 分割: 2024_01_07T21_59_12.253105
- 路径: **/details_harness|gsm8k|5_2024-01-07T21-59-12.253105.parquet
- 分割: 2024_01_08T03_36_26.320528
- 路径: **/details_harness|gsm8k|5_2024-01-08T03-36-26.320528.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2024-01-08T03-36-26.320528.parquet
- 分割: 2024_01_07T21_59_12.253105
- 数据文件:
-
配置名称: harness_hellaswag_10
- 数据文件:
- 分割: 2024_01_07T21_59_12.253105
- 路径: **/details_harness|hellaswag|10_2024-01-07T21-59-12.253105.parquet
- 分割: 2024_01_08T03_36_26.320528
- 路径: **/details_harness|hellaswag|10_2024-01-08T03-36-26.320528.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2024-01-08T03-36-26.320528.parquet
- 分割: 2024_01_07T21_59_12.253105
- 数据文件:
-
配置名称: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_01_07T21_59_12.253105
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-human_aging|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-international_law|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-management|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-marketing|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2024-01-07T21-59-12.253105.parquet
- **/details_harness|hendrycksTest-nutrition|5_2024-01-07T21-59-12.253105.
- 路径:
- 分割: 2024_01_07T21_59_12.253105
- 数据文件:



