open-llm-leaderboard/details_KoboldAI__OPT-6B-nerys-v2
收藏数据集概述
数据集简介
该数据集是在对模型 KoboldAI/OPT-6B-nerys-v2 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。
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。
- 创建来源:数据集来自2次运行,每次运行在每个配置中作为一个特定的分片存在,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:一个额外的配置 "results" 存储所有运行的汇总结果,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_KoboldAI__OPT-6B-nerys-v2", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-17T17:27:37.555726 运行的最新结果: python { "all": { "em": 0.001363255033557047, "em_stderr": 0.00037786091964606383, "f1": 0.04780515939597329, "f1_stderr": 0.0011803538106248462, "acc": 0.331351691297071, "acc_stderr": 0.007809829503630222 }, "harness|drop|3": { "em": 0.001363255033557047, "em_stderr": 0.00037786091964606383, "f1": 0.04780515939597329, "f1_stderr": 0.0011803538106248462 }, "harness|gsm8k|5": { "acc": 0.006823351023502654, "acc_stderr": 0.0022675371022545087 }, "harness|winogrande|5": { "acc": 0.6558800315706393, "acc_stderr": 0.013352121905005935 } }
配置详情
-
harness_arc_challenge_25
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-19T15:44:43.030305.parquet
-
harness_drop_3
- 分片:2023_10_17T17_27_37.555726, latest
- 路径:
**/details_harness|drop|3_2023-10-17T17-27-37.555726.parquet
-
harness_gsm8k_5
- 分片:2023_10_17T17_27_37.555726, latest
- 路径:
**/details_harness|gsm8k|5_2023-10-17T17-27-37.555726.parquet
-
harness_hellaswag_10
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hellaswag|10_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_anatomy_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_astronomy_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_business_ethics_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_college_biology_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_college_medicine_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_college_physics_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_computer_security_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_econometrics_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T15:44:43.030305.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分片:2023_07_19T15_44_43.030305, latest
- 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-07-19T15:44:43.030305.parquet
以上是对数据集的详细概述,包括数据集的简介、结构、加载示例、最新结果以及各个配置的详细信息。



