open-llm-leaderboard/details_PSanni__Deer-3b
收藏数据集概述
数据集名为 "Evaluation run of PSanni/Deer-3b",是在对模型 PSanni/Deer-3b 进行评估运行期间自动创建的。该数据集用于 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 Open LLM Leaderboard 上的聚合指标。
数据加载示例
以下是加载特定运行详细信息的示例代码:
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_PSanni__Deer-3b", "harness_winogrande_5", split="train")
最新结果
以下是 2023-09-16T20:50:46.284611 运行 的最新结果:
python { "all": { "em": 0.0003145973154362416, "em_stderr": 0.0001816137946883968, "f1": 0.04833053691275181, "f1_stderr": 0.0011657715269814616, "acc": 0.28880911790700303, "acc_stderr": 0.0077049156139354594 }, "harness|drop|3": { "em": 0.0003145973154362416, "em_stderr": 0.0001816137946883968, "f1": 0.04833053691275181, "f1_stderr": 0.0011657715269814616 }, "harness|gsm8k|5": { "acc": 0.003032600454890068, "acc_stderr": 0.0015145735612245434 }, "harness|winogrande|5": { "acc": 0.574585635359116, "acc_stderr": 0.013895257666646375 } }
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运行时间戳。以下是部分配置的详细信息:
-
harness_arc_challenge_25
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|arc:challenge|25_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-09T14:13:49.318775.parquet
-
harness_drop_3
- 分割: 2023_09_16T20_50_46.284611
- 路径:
**/details_harness|drop|3_2023-09-16T20-50-46.284611.parquet - 分割: latest
- 路径:
**/details_harness|drop|3_2023-09-16T20-50-46.284611.parquet
-
harness_gsm8k_5
- 分割: 2023_09_16T20_50_46.284611
- 路径:
**/details_harness|gsm8k|5_2023-09-16T20-50-46.284611.parquet - 分割: latest
- 路径:
**/details_harness|gsm8k|5_2023-09-16T20-50-46.284611.parquet
-
harness_hellaswag_10
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hellaswag|10_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hellaswag|10_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T14:13:49.318775.parquet等 - 分割: latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T14:13:49.318775.parquet等
-
harness_hendrycksTest_abstract_algebra_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_anatomy_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_astronomy_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_college_biology_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_college_physics_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_computer_security_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_econometrics_5
- 分割: 2023_08_09T14_13_49.318775
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-09T14:13:49.318775.parquet - 分割: latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-09T14:13:49.318775.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割: 2023_08_09T14_13_49.318775
- 路径: `**/details_harness|hendrycksTest-electrical_engineering



