open-llm-leaderboard-old/details_OpenLemur__lemur-70b-chat-v1
收藏数据集概述
数据集简介
该数据集是在评估模型 OpenLemur/lemur-70b-chat-v1 在 Open LLM Leaderboard 上的自动创建的。
数据集结构
- 配置数量:64个配置,每个配置对应一个评估任务。
- 数据来源:数据集由2次运行结果组成,每次运行的结果作为特定分片存储,分片名称使用运行的时间戳。
- 最新结果:"train" 分片始终指向最新的结果。
- 汇总结果:额外配置 "results" 存储所有运行的汇总结果,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_OpenLemur__lemur-70b-chat-v1", "harness_winogrande_5", split="train")
最新结果
以下是 2023-09-17T13:31:04.707005 运行的最新结果: python { "all": { "em": 0.006711409395973154, "em_stderr": 0.0008361500895152445, "f1": 0.0739702181208053, "f1_stderr": 0.001585201628872726, "acc": 0.5850941225115532, "acc_stderr": 0.01201805791264202 }, "harness|drop|3": { "em": 0.006711409395973154, "em_stderr": 0.0008361500895152445, "f1": 0.0739702181208053, "f1_stderr": 0.001585201628872726 }, "harness|gsm8k|5": { "acc": 0.35329795299469297, "acc_stderr": 0.013166337192115683 }, "harness|winogrande|5": { "acc": 0.8168902920284136, "acc_stderr": 0.010869778633168358 } }
配置详情
以下是数据集的部分配置详情:
-
harness_arc_challenge_25
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-24T04:11:57.870589.parquet
-
harness_drop_3
- 分片:2023_09_17T13_31_04.707005, latest
- 路径:
**/details_harness|drop|3_2023-09-17T13-31-04.707005.parquet
-
harness_gsm8k_5
- 分片:2023_09_17T13_31_04.707005, latest
- 路径:
**/details_harness|gsm8k|5_2023-09-17T13-31-04.707005.parquet
-
harness_hellaswag_10
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hellaswag|10_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_anatomy_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_astronomy_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_business_ethics_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_college_biology_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_college_medicine_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_college_physics_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_computer_security_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_econometrics_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-24T04:11:57.870589.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分片:2023_08_24T04_11_57.870589, latest
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-24T04:11:57.870589.parquet



