open-llm-leaderboard/details_OpenBuddy__openbuddy-mistral-7b-v13.1
收藏数据集概述
数据集名称
- Evaluation run of OpenBuddy/openbuddy-mistral-7b-v13.1
数据集描述
- 该数据集是在模型 OpenBuddy/openbuddy-mistral-7b-v13.1 在 Open LLM Leaderboard 上的评估运行期间自动创建的。
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作为一个特定的分片找到,分片名称使用运行的时间戳。
- "train" 分片始终指向最新的结果。
- 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_OpenBuddy__openbuddy-mistral-7b-v13.1", "harness_winogrande_5", split="train")
最新结果
- 这些是最新结果,来自 2023-10-27T07:58:51.376578 的运行: python { "all": { "em": 0.28953439597315433, "em_stderr": 0.004644738434561709, "f1": 0.3589429530201351, "f1_stderr": 0.004562237952673667, "acc": 0.401525754664538, "acc_stderr": 0.010223042101778138 }, "harness|drop|3": { "em": 0.28953439597315433, "em_stderr": 0.004644738434561709, "f1": 0.3589429530201351, "f1_stderr": 0.004562237952673667 }, "harness|gsm8k|5": { "acc": 0.08718726307808947, "acc_stderr": 0.007770691416783547 }, "harness|winogrande|5": { "acc": 0.7158642462509865, "acc_stderr": 0.01267539278677273 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|arc:challenge|25_2023-10-11T17-55-12.784881.parquet
-
harness_drop_3
- 分片:2023_10_27T07_58_51.376578, latest
- 路径:
**/details_harness|drop|3_2023-10-27T07-58-51.376578.parquet
-
harness_gsm8k_5
- 分片:2023_10_27T07_58_51.376578, latest
- 路径:
**/details_harness|gsm8k|5_2023-10-27T07-58-51.376578.parquet
-
harness_hellaswag_10
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hellaswag|10_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_anatomy_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_astronomy_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_business_ethics_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_college_biology_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_college_medicine_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_college_physics_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_computer_security_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_econometrics_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-11T17-55-12.784881.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分片:2023_10_11T17_55_12.784881, latest
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-11T17-55-12.784881.parquet
以上是数据集的详细概述,包括数据集名称、描述、组成、加载示例、最新结果以及各个配置的详细信息。



