open-llm-leaderboard-old/details_Riiid__sheep-duck-llama-2-70b-v1.1
收藏数据集概述
数据集简介
该数据集是在模型Riiid/sheep-duck-llama-2-70b-v1.1的评估运行期间自动创建的,用于Open LLM Leaderboard。
数据集结构
- 配置数量:63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,每个运行在每个配置中作为一个特定的分割存在,分割名称使用运行的时间戳。
- 训练分割:"train"分割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,用于计算和显示聚合指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Riiid__sheep-duck-llama-2-70b-v1.1", "harness_winogrande_5", split="train")
最新结果
以下是最新结果: python { "all": { "acc": 0.7053343320815155, "acc_stderr": 0.030260160101824644, "acc_norm": 0.7109334613998801, "acc_norm_stderr": 0.03084136530304881, "mc1": 0.4663402692778458, "mc1_stderr": 0.017463793867168103, "mc2": 0.6454578975757769, "mc2_stderr": 0.014741040304266572 }, "harness|arc:challenge|25": { "acc": 0.6868600682593856, "acc_stderr": 0.013552671543623501, "acc_norm": 0.7312286689419796, "acc_norm_stderr": 0.012955065963710688 }, "harness|hellaswag|10": { "acc": 0.691396136227843, "acc_stderr": 0.004609731925736905, "acc_norm": 0.8777136028679546, "acc_norm_stderr": 0.0032694673590543157 }, # 其他任务的结果... }
配置详情
-
harness_arc_challenge_25
- 分割:2023_10_04T07_20_47.805368, 2023_12_09T22_48_41.234684, latest
- 路径:
**/details_harness|arc:challenge|25_2023-10-04T07-20-47.805368.parquet,**/details_harness|arc:challenge|25_2023-12-09T22-48-41.234684.parquet
-
harness_gsm8k_5
- 分割:2023_12_09T22_48_41.234684, latest
- 路径:
**/details_harness|gsm8k|5_2023-12-09T22-48-41.234684.parquet
-
harness_hellaswag_10
- 分割:2023_10_04T07_20_47.805368, 2023_12_09T22_48_41.234684, latest
- 路径:
**/details_harness|hellaswag|10_2023-10-04T07-20-47.805368.parquet,**/details_harness|hellaswag|10_2023-12-09T22-48-41.234684.parquet
-
harness_hendrycksTest_5
- 分割:2023_10_04T07_20_47.805368, 2023_12_09T22_48_41.234684, latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-anatomy|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-astronomy|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-business_ethics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-college_biology|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-college_medicine|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-college_physics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-computer_security|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-econometrics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-formal_logic|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-global_facts|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-human_aging|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-international_law|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-machine_learning|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-management|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-marketing|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-nutrition|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-philosophy|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-prehistory|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-professional_law|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-professional_medicine|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-professional_psychology|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-public_relations|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-security_studies|5_2023-10-04T07-20-47.805368.parquet,**/details_harness|hendrycksTest-sociology|5_2023-10-04T07-20-47.805368.parquet, `**/details_harness|hendrycksTest-us_foreign_policy|5_2023-10-



