open-llm-leaderboard-old/details_openchat__openchat_v2
收藏数据集概述
数据集名称
- Evaluation run of openchat/openchat_v2
数据集组成
- 该数据集包含64个配置,每个配置对应一个评估任务。
- 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。
- "train"分片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_openchat__openchat_v2", "harness_winogrande_5", split="train")
最新结果
- 最新结果来自2023-10-18T23:33:59.473281运行,包含以下指标: python { "all": { "em": 0.0016778523489932886, "em_stderr": 0.0004191330178826953, "f1": 0.06369546979865812, "f1_stderr": 0.0013881754743750058, "acc": 0.4267044764366107, "acc_stderr": 0.009941310874908384 }, "harness|drop|3": { "em": 0.0016778523489932886, "em_stderr": 0.0004191330178826953, "f1": 0.06369546979865812, "f1_stderr": 0.0013881754743750058 }, "harness|gsm8k|5": { "acc": 0.09097801364670205, "acc_stderr": 0.007921322844013628 }, "harness|winogrande|5": { "acc": 0.7624309392265194, "acc_stderr": 0.011961298905803141 } }
配置详情
-
harness_arc_challenge_25
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|arc:challenge|25_2023-07-24T16:15:43.375202.parquet
-
harness_drop_3
- 分片: 2023_10_18T23_33_59.473281, latest
- 路径:
**/details_harness|drop|3_2023-10-18T23-33-59.473281.parquet
-
harness_gsm8k_5
- 分片: 2023_10_18T23_33_59.473281, latest
- 路径:
**/details_harness|gsm8k|5_2023-10-18T23-33-59.473281.parquet
-
harness_hellaswag_10
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hellaswag|10_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径: 多个文件路径,例如
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_abstract_algebra_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_anatomy_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_astronomy_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_business_ethics_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_clinical_knowledge_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_college_biology_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_college_chemistry_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_college_computer_science_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_college_mathematics_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_college_medicine_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_college_physics_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_computer_security_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_conceptual_physics_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_econometrics_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_electrical_engineering_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-24T16:15:43.375202.parquet
-
harness_hendrycksTest_elementary_mathematics_5
- 分片: 2023_07_24T16_15_43.375202, latest
- 路径: `**/details_harness|hendrycksTest-elementary_mathematics|5_2023



