open-llm-leaderboard/details_maximuslee07__llama-2-7b-rockwell-final
收藏数据集概述
该数据集是在对模型 maximuslee07/llama-2-7b-rockwell-final 进行评估运行期间自动创建的。数据集包含64个配置,每个配置对应一个评估任务。数据集从2次运行中创建,每个运行可以在每个配置中作为一个特定的分片找到,分片名称使用运行的时间戳。"train" 分片始终指向最新的结果。
数据集结构
数据集包含以下配置:
-
harness_arc_challenge_25
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|arc:challenge|25_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|arc:challenge|25_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_drop_3
- 分片:
2023_10_23T20_01_16.627369 - 路径:
**/details_harness|drop|3_2023-10-23T20-01-16.627369.parquet - 分片:
latest - 路径:
**/details_harness|drop|3_2023-10-23T20-01-16.627369.parquet
- 分片:
-
harness_gsm8k_5
- 分片:
2023_10_23T20_01_16.627369 - 路径:
**/details_harness|gsm8k|5_2023-10-23T20-01-16.627369.parquet - 分片:
latest - 路径:
**/details_harness|gsm8k|5_2023-10-23T20-01-16.627369.parquet
- 分片:
-
harness_hellaswag_10
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hellaswag|10_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hellaswag|10_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-anatomy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-astronomy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_biology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_physics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-computer_security|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-econometrics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-global_facts|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-human_aging|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-international_law|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-management|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-marketing|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-nutrition|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-philosophy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-prehistory|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_law|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-public_relations|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-security_studies|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-sociology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-virology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-world_religions|5_2023-10-04T01-33-36.813954.parquet
- 分片:
latest - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-anatomy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-astronomy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_biology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-college_physics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-computer_security|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-econometrics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-global_facts|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-human_aging|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-international_law|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-management|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-marketing|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-nutrition|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-philosophy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-prehistory|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_law|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-public_relations|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-security_studies|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-sociology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-virology|5_2023-10-04T01-33-36.813954.parquet**/details_harness|hendrycksTest-world_religions|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_abstract_algebra_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_anatomy_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-anatomy|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_astronomy_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-astronomy|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_business_ethics_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_clinical_knowledge_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_college_biology_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-college_biology|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_college_chemistry_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_college_computer_science_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_college_mathematics_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_college_medicine_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_college_physics_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-college_physics|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_computer_security_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-computer_security|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_conceptual_physics_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_econometrics_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-econometrics|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_electrical_engineering_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T01-33-36.813954.parquet
- 分片:
-
harness_hendrycksTest_elementary_mathematics_5
- 分片:
2023_10_04T01_33_36.813954 - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T01-33-36.813954.parquet - 分片:
latest - 路径:
**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T01-33-36.813954.parquet
- 分片:
数据集加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_maximuslee07__llama-2-7b-rockwell-final", "harness_winogrande_5", split="train")
最新结果
这些是最新结果(来自2023-10-23T20:01:16.627369)的示例:
python { "all": { "em": 0.03460570469798658, "em_stderr": 0.001871827675399587, "f1": 0.100573615771812, "f1_stderr": 0.002343392042876464, "acc": 0.3819496844432025, "acc_stderr": 0.010259509540838537 }, "harness|drop|3": { "em": 0.03460570469798658, "em_stderr": 0.001871827675399587, "f1": 0.100573615771812, "f1_stderr": 0.002343392042876464 }, "harness|gsm8k|5": { "acc": 0.07960576194086429, "acc_stderr": 0.007455924338676263 }, "harness|winogrande|5": { "acc": 0.6842936069455406, "acc_stderr": 0.01306309474300081 } }



