open-llm-leaderboard-old/details_bofenghuang__vigogne-13b-chat
收藏数据集概述
该数据集是在评估模型 bofenghuang/vigogne-13b-chat 在 Open LLM Leaderboard 上的自动创建的。数据集包含 64 个配置,每个配置对应一个评估任务。
数据集结构
- 配置数量:64 个配置
- 创建来源:从 2 次运行中创建,每个运行在每个配置中作为一个特定的分割存在,分割名称使用运行的时间戳。
- 训练分割:"train" 分割始终指向最新的结果。
- 结果配置:一个额外的配置 "results" 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_bofenghuang__vigogne-13b-chat", "harness_winogrande_5", split="train")
最新结果
以下是 2023-09-22T17:13:01.877874 运行的最新结果:
python { "all": { "em": 0.28093540268456374, "em_stderr": 0.004602850194300876, "f1": 0.3580620805369141, "f1_stderr": 0.0045122087324351344, "acc": 0.42528152381590656, "acc_stderr": 0.009746925675481622 }, "harness|drop|3": { "em": 0.28093540268456374, "em_stderr": 0.004602850194300876, "f1": 0.3580620805369141, "f1_stderr": 0.0045122087324351344 }, "harness|gsm8k|5": { "acc": 0.08339651250947688, "acc_stderr": 0.007615650277106699 }, "harness|winogrande|5": { "acc": 0.7671665351223362, "acc_stderr": 0.011878201073856544 } }
配置详情
-
harness_arc_challenge_25
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-25T11:18:55.409320.parquet
-
harness_drop_3
- 分割:2023_09_22T17_13_01.877874, latest
- 路径:
**/details_harness|drop|3_2023-09-22T17-13-01.877874.parquet
-
harness_gsm8k_5
- 分割:2023_09_22T17_13_01.877874, latest
- 路径:
**/details_harness|gsm8k|5_2023-09-22T17-13-01.877874.parquet
-
harness_hellaswag_10
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hellaswag|10_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-25T11:18:55.409320.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分割:2023_07_25T11_18_55.409320, latest
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-25T11:18:55.409320.parquet
以上是数据集的详细概述,包括数据集的结构、加载示例、最新结果以及各个配置的详细信息。



