open-llm-leaderboard/details_roneneldan__TinyStories-1M
收藏数据集概述
数据集 Evaluation run of roneneldan/TinyStories-1M 是在评估模型 roneneldan/TinyStories-1M 在 Open LLM Leaderboard 上的自动创建的。
数据集组成
- 数据集包含 64 个配置,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。
加载数据
可以使用以下代码加载特定运行的详细信息:
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_roneneldan__TinyStories-1M", "harness_winogrande_5", split="train")
最新结果
以下是 2023-09-22T21:41:24.294253 运行的最新结果:
python { "all": { "em": 0.00020973154362416107, "em_stderr": 0.00014829481977282063, "f1": 0.003178481543624158, "f1_stderr": 0.0002730192207643319, "acc": 0.26085240726124703, "acc_stderr": 0.007019619608242314 }, "harness|drop|3": { "em": 0.00020973154362416107, "em_stderr": 0.00014829481977282063, "f1": 0.003178481543624158, "f1_stderr": 0.0002730192207643319 }, "harness|gsm8k|5": { "acc": 0.0, "acc_stderr": 0.0 }, "harness|winogrande|5": { "acc": 0.5217048145224941, "acc_stderr": 0.014039239216484627 } }
配置详情
数据集包含多个配置,每个配置对应不同的任务和数据文件路径。以下是部分配置示例:
-
harness_arc_challenge_25
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T13:25:02.593147.parquet
-
harness_drop_3
- 分割:2023_09_22T21_41_24.294253
- 路径:
**/details_harness|drop|3_2023-09-22T21-41-24.294253.parquet - 分割:latest
- 路径:
**/details_harness|drop|3_2023-09-22T21-41-24.294253.parquet
-
harness_gsm8k_5
- 分割:2023_09_22T21_41_24.294253
- 路径:
**/details_harness|gsm8k|5_2023-09-22T21-41-24.294253.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-09-22T21-41-24.294253.parquet
-
harness_hellaswag_10
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hellaswag|10_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T13:25:02.593147.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_07_19T13_25_02.593147
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T13:25:02.593147.parquet - 分割:latest
- 路径:`**/details_har



