open-llm-leaderboard-old/details_AA051615__A0304
收藏数据集概述
数据集描述
该数据集是在模型AA051615/A0304在Open LLM Leaderboard上的评估运行期间自动创建的。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
- 配置数量:63个配置
- 数据来源:从1次运行中创建,每个运行在每个配置中作为一个特定的分片存在,分片名称使用运行的时间戳。
- 分片命名:使用运行的时间戳命名
- 最新结果:"train"分片始终指向最新结果
额外配置
- 结果配置:名为"results"的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_AA051615__A0304", "harness_winogrande_5", split="train")
最新结果
以下是2024-03-05T02:38:53.174033运行的最新结果:
python { "all": { "acc": 0.8329894986297788, "acc_stderr": 0.0242885840106617, "acc_norm": 0.8418851861297316, "acc_norm_stderr": 0.024662491905159404, "mc1": 0.3659730722154223, "mc1_stderr": 0.016862941684088365, "mc2": 0.5334887043569735, "mc2_stderr": 0.01542333936417327 }, "harness|arc:challenge|25": { "acc": 0.6296928327645052, "acc_stderr": 0.01411129875167495, "acc_norm": 0.6757679180887372, "acc_norm_stderr": 0.01367881039951882 }, "harness|hellaswag|10": { "acc": 0.6292571200955985, "acc_stderr": 0.0048201660022530795, "acc_norm": 0.8278231428002389, "acc_norm_stderr": 0.003767625141611702 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.64, "acc_stderr": 0.048241815132442176, "acc_norm": 0.64, "acc_norm_stderr": 0.048241815132442176 }, # 其他任务的结果... }
配置详情
-
配置名称:harness_arc_challenge_25
- 数据文件:
- 分片:2024_03_05T02_38_53.174033
- 路径:
**/details_harness|arc:challenge|25_2024-03-05T02-38-53.174033.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|arc:challenge|25_2024-03-05T02-38-53.174033.parquet
- 路径:
- 分片:2024_03_05T02_38_53.174033
- 数据文件:
-
配置名称:harness_gsm8k_5
- 数据文件:
- 分片:2024_03_05T02_38_53.174033
- 路径:
**/details_harness|gsm8k|5_2024-03-05T02-38-53.174033.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|gsm8k|5_2024-03-05T02-38-53.174033.parquet
- 路径:
- 分片:2024_03_05T02_38_53.174033
- 数据文件:
-
配置名称:harness_hellaswag_10
- 数据文件:
- 分片:2024_03_05T02_38_53.174033
- 路径:
**/details_harness|hellaswag|10_2024-03-05T02-38-53.174033.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|hellaswag|10_2024-03-05T02-38-53.174033.parquet
- 路径:
- 分片:2024_03_05T02_38_53.174033
- 数据文件:
-
配置名称:harness_hendrycksTest_5
- 数据文件:
- 分片:2024_03_05T02_38_53.174033
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-anatomy|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-astronomy|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-college_biology|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-college_physics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-computer_security|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-econometrics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-global_facts|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_geography|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_physics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-human_aging|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-human_sexuality|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-international_law|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-jurisprudence|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-machine_learning|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-management|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-marketing|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-medical_genetics|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-miscellaneous|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-moral_disputes|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-nutrition|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-philosophy|5_2024-03-05T02-38-53.174033.parquet**/details_harness|hendrycksTest-prehistory|5_2024-03-05T02-38-53.174033.parquet- `**/details_harness|hendrycksTest
- 路径:
- 分片:2024_03_05T02_38_53.174033
- 数据文件:



