open-llm-leaderboard/details_togethercomputer__GPT-JT-6B-v0
收藏数据集概述
该数据集是在对模型 togethercomputer/GPT-JT-6B-v0 进行评估运行期间自动创建的。数据集包含 64 个配置,每个配置对应一个评估任务。数据集从 2 次运行中创建,每个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。
数据集结构
配置列表
-
harness_arc_challenge_25
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|arc:challenge|25_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|arc:challenge|25_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_drop_3
- 分割:
2023_10_17T19_26_54.220051 - 路径:
**/details_harness|drop|3_2023-10-17T19-26-54.220051.parquet - 分割:
latest - 路径:
**/details_harness|drop|3_2023-10-17T19-26-54.220051.parquet
- 分割:
-
harness_gsm8k_5
- 分割:
2023_10_17T19_26_54.220051 - 路径:
**/details_harness|gsm8k|5_2023-10-17T19-26-54.220051.parquet - 分割:
latest - 路径:
**/details_harness|gsm8k|5_2023-10-17T19-26-54.220051.parquet
- 分割:
-
harness_hellaswag_10
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hellaswag|10_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hellaswag|10_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:42:14.994932.parquet**/details_harness|hendrycksTest-anatomy|5_2023-07-19T15:42:14.994932.parquet- ... (其他路径省略)
- 分割:
latest - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:42:14.994932.parquet**/details_harness|hendrycksTest-anatomy|5_2023-07-19T15:42:14.994932.parquet- ... (其他路径省略)
- 分割:
-
harness_hendrycksTest_abstract_algebra_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_anatomy_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_astronomy_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_business_ethics_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_college_biology_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_college_chemistry_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_college_computer_science_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_college_mathematics_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_college_medicine_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_college_physics_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_computer_security_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_conceptual_physics_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_econometrics_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_electrical_engineering_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T15:42:14.994932.parquet
- 分割:
-
harness_hendrycksTest_elementary_mathematics_5
- 分割:
2023_07_19T15_42_14.994932 - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T15:42:14.994932.parquet - 分割:
latest - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T15:42:14.994932.parquet
- 分割:
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_togethercomputer__GPT-JT-6B-v0", "harness_winogrande_5", split="train")
最新结果
以下是最新结果(来自运行 2023-10-17T19:26:54.220051)的部分数据:
python { "all": { "em": 0.0010486577181208054, "em_stderr": 0.0003314581465219154, "f1": 0.043061031879194765, "f1_stderr": 0.0011437900819203201, "acc": 0.330058886781919, "acc_stderr": 0.008219084533910332 }, "harness|drop|3": { "em": 0.0010486577181208054, "em_stderr": 0.000



