open-llm-leaderboard-old/details_Voicelab__trurl-2-7b
收藏数据集概述
该数据集是在评估模型 Voicelab/trurl-2-7b 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
- 创建来源:从 2 次运行中创建,每次运行对应一个特定的时间戳
- 数据分割:每个配置包含多个分割,其中 "train" 分割指向最新的结果
- 额外配置:"results" 配置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_Voicelab__trurl-2-7b", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-24T13:00:35.734451 运行的最新结果:
python { "all": { "em": 0.26908557046979864, "em_stderr": 0.004541696656496853, "f1": 0.3290079697986583, "f1_stderr": 0.004499453214736992, "acc": 0.3967222424009962, "acc_stderr": 0.009837690155913053 }, "harness|drop|3": { "em": 0.26908557046979864, "em_stderr": 0.004541696656496853, "f1": 0.3290079697986583, "f1_stderr": 0.004499453214736992 }, "harness|gsm8k|5": { "acc": 0.0712661106899166, "acc_stderr": 0.007086462127954499 }, "harness|winogrande|5": { "acc": 0.7221783741120757, "acc_stderr": 0.012588918183871605 } }
配置详情
-
harness_arc_challenge_25
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|arc:challenge|25_2023-08-17T14:14:32.422343.parquet
-
harness_drop_3
- 分割:2023_10_24T13_00_35.734451, latest
- 路径:
**/details_harness|drop|3_2023-10-24T13-00-35.734451.parquet
-
harness_gsm8k_5
- 分割:2023_10_24T13_00_35.734451, latest
- 路径:
**/details_harness|gsm8k|5_2023-10-24T13-00-35.734451.parquet
-
harness_hellaswag_10
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hellaswag|10_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-08-17T14:14:32.422343.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分割:2023_08_17T14_14_32.422343, latest
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-17T14:14:32.422343.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。



