open-llm-leaderboard-old/details_Phind__Phind-CodeLlama-34B-Python-v1
收藏数据集概述
数据集简介
该数据集是在评估模型Phind/Phind-CodeLlama-34B-Python-v1在Open LLM Leaderboard上的运行过程中自动创建的。
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。
- 运行次数: 数据集由2次运行创建,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。
- 训练分片: "train"分片始终指向最新的结果。
- 结果配置: 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
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Phind__Phind-CodeLlama-34B-Python-v1", "harness_winogrande_5", split="train")
最新结果
以下是2023-09-17T16:02:38.595550运行的最新结果: python { "all": { "em": 0.265625, "em_stderr": 0.004523067479107055, "f1": 0.3185192953020138, "f1_stderr": 0.004482746835839152, "acc": 0.4517772845779581, "acc_stderr": 0.012170333746109104 }, "harness|drop|3": { "em": 0.265625, "em_stderr": 0.004523067479107055, "f1": 0.3185192953020138, "f1_stderr": 0.004482746835839152 }, "harness|gsm8k|5": { "acc": 0.21531463229719486, "acc_stderr": 0.011322096294579654 }, "harness|winogrande|5": { "acc": 0.6882399368587214, "acc_stderr": 0.013018571197638551 } }
配置详情
-
harness_arc_challenge_25
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|arc:challenge|25_2023-08-26T05:45:26.681000.parquet
-
harness_drop_3
- 分片: 2023_09_17T16_02_38.595550, latest
- 路径: **/details_harness|drop|3_2023-09-17T16-02-38.595550.parquet
-
harness_gsm8k_5
- 分片: 2023_09_17T16_02_38.595550, latest
- 路径: **/details_harness|gsm8k|5_2023-09-17T16-02-38.595550.parquet
-
harness_hellaswag_10
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hellaswag|10_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: 多个路径,包括但不限于 **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_anatomy_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-anatomy|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_astronomy_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-astronomy|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_business_ethics_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-business_ethics|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_college_biology_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-college_biology|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-college_chemistry|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-college_computer_science|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-college_mathematics|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_college_medicine_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-college_medicine|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_college_physics_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-college_physics|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_computer_security_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-computer_security|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_econometrics_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-econometrics|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-26T05:45:26.681000.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分片: 2023_08_26T05_45_26.681000, latest
- 路径: **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-26T05:45:26.681000.parquet
以上是数据集的详细概述,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。



