open-llm-leaderboard-old/details_migtissera__Synthia-70B-v1.2
收藏数据集概述
该数据集是在对模型 migtissera/Synthia-70B-v1.2 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。
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。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_migtissera__Synthia-70B-v1.2", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-17T14:51:20.480254 运行 的最新结果:
python { "all": { "em": 0.364618288590604, "em_stderr": 0.004929197624393639, "f1": 0.42417365771812215, "f1_stderr": 0.004776577842624861, "acc": 0.5759284047791582, "acc_stderr": 0.011665477241539865 }, "harness|drop|3": { "em": 0.364618288590604, "em_stderr": 0.004929197624393639, "f1": 0.42417365771812215, "f1_stderr": 0.004776577842624861 }, "harness|gsm8k|5": { "acc": 0.3191811978771797, "acc_stderr": 0.012840345676251651 }, "harness|winogrande|5": { "acc": 0.8326756116811366, "acc_stderr": 0.010490608806828079 } }
配置详情
-
harness_arc_challenge_25
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|arc:challenge|25_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-09-02T17:59:05.420313.parquet
-
harness_drop_3
- 分割:2023_10_17T14_51_20.480254
- 路径:
**/details_harness|drop|3_2023-10-17T14-51-20.480254.parquet - 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-17T14-51-20.480254.parquet
-
harness_gsm8k_5
- 分割:2023_10_17T14_51_20.480254
- 路径:
**/details_harness|gsm8k|5_2023-10-17T14-51-20.480254.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-17T14-51-20.480254.parquet
-
harness_hellaswag_10
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hellaswag|10_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-02T17:59:05.420313.parquet等 42 个文件 - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-02T17:59:05.420313.parquet等 42 个文件
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-09-02T17:59:05.420313.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_09_02T17_59_05.420313
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-02T17:59:05.420313.parquet - 分割:latest
- 路径:`**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-



