open-llm-leaderboard/details_hiyouga__Baichuan2-7B-Chat-LLaMAfied
收藏数据集概述
数据集简介
该数据集是在评估模型 hiyouga/Baichuan2-7B-Chat-LLaMAfied 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个配置
- 数据来源:从2次运行中创建,每次运行对应一个特定的时间戳分割。
- 最新结果:"train" 分割始终指向最新结果。
- 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_hiyouga__Baichuan2-7B-Chat-LLaMAfied", "harness_winogrande_5", split="train")
最新结果
以下是 最新结果来自 run 2023-10-27T01:55:17.464897 的摘要: python { "all": { "em": 0.2062709731543624, "em_stderr": 0.004143762363131985, "f1": 0.26938129194630883, "f1_stderr": 0.004172682699820514, "acc": 0.40028530858265426, "acc_stderr": 0.010786124750718863 }, "harness|drop|3": { "em": 0.2062709731543624, "em_stderr": 0.004143762363131985, "f1": 0.26938129194630883, "f1_stderr": 0.004172682699820514 }, "harness|gsm8k|5": { "acc": 0.10917361637604246, "acc_stderr": 0.008590089300511151 }, "harness|winogrande|5": { "acc": 0.691397000789266, "acc_stderr": 0.012982160200926577 } }
配置详情
以下是部分配置的详细信息:
-
harness_arc_challenge_25
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|arc:challenge|25_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-10-10T07-31-02.024016.parquet
-
harness_drop_3
- 分割:2023_10_27T01_55_17.464897
- 路径:
**/details_harness|drop|3_2023-10-27T01-55-17.464897.parquet - 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-27T01-55-17.464897.parquet
-
harness_gsm8k_5
- 分割:2023_10_27T01_55_17.464897
- 路径:
**/details_harness|gsm8k|5_2023-10-27T01-55-17.464897.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-27T01-55-17.464897.parquet
-
harness_hellaswag_10
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hellaswag|10_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_5
- 分割:2023_10_10T07_31_02.024016
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:多个路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_10_10T07_31_02.024016
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-10-10T07-31-02.024016.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-10-10T07-31-02.024016.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_10_10T07_31_02.024016
- 路径:`**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-10T07-31-02.0



