open-llm-leaderboard/details_Rallio67__3B-redpajama-conditional-alpha
收藏数据集概述
该数据集是在评估模型 Rallio67/3B-redpajama-conditional-alpha 在 Open LLM Leaderboard 上的自动创建的。
数据集组成
- 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,每次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。
- 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_Rallio67__3B-redpajama-conditional-alpha", "harness_winogrande_5", split="train")
最新结果
以下是 2023-10-28T09:28:50.380809 运行的最新结果:
python { "all": { "em": 0.0036703020134228187, "em_stderr": 0.0006192871806511201, "f1": 0.052849203020134294, "f1_stderr": 0.00137566520570553, "acc": 0.3069000037698072, "acc_stderr": 0.007930535381730674 }, "harness|drop|3": { "em": 0.0036703020134228187, "em_stderr": 0.0006192871806511201, "f1": 0.052849203020134294, "f1_stderr": 0.00137566520570553 }, "harness|gsm8k|5": { "acc": 0.006065200909780136, "acc_stderr": 0.002138670301460461 }, "harness|winogrande|5": { "acc": 0.6077348066298343, "acc_stderr": 0.013722400462000888 } }
配置详情
-
harness_arc_challenge_25
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|arc:challenge|25_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2023-09-18T14-36-57.601576.parquet
-
harness_drop_3
- 分割:2023_10_28T09_28_50.380809
- 路径:
**/details_harness|drop|3_2023-10-28T09-28-50.380809.parquet - 分割:latest
- 路径:
**/details_harness|drop|3_2023-10-28T09-28-50.380809.parquet
-
harness_gsm8k_5
- 分割:2023_10_28T09_28_50.380809
- 路径:
**/details_harness|gsm8k|5_2023-10-28T09-28-50.380809.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2023-10-28T09-28-50.380809.parquet
-
harness_hellaswag_10
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hellaswag|10_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-18T14-36-57.601576.parquet等 40 个文件 - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-18T14-36-57.601576.parquet等 40 个文件
-
harness_hendrycksTest_abstract_algebra_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_anatomy_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_astronomy_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_business_ethics_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_college_biology_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_college_medicine_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_college_physics_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_computer_security_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_econometrics_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-09-18T14-36-57.601576.parquet - 分割:latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-09-18T14-36-57.601576.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分割:2023_09_18T14_36_57.601576
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-09-18T14-36-57.601576.parquet



