open-llm-leaderboard-old/details_rwitz2__ipo-test
收藏数据集概述
数据集来源
该数据集是在对模型 rwitz2/ipo-test 进行评估运行期间自动创建的,评估运行在 Open LLM Leaderboard 上进行。
数据集组成
数据集由 63 个配置组成,每个配置对应一个评估任务。数据集从 1 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。
结果配置
一个额外的配置 "results" 存储了所有运行的聚合结果,用于计算并在 Open LLM Leaderboard 上显示聚合指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_rwitz2__ipo-test", "harness_winogrande_5", split="train")
最新结果
以下是 2023-12-12T03:53:21.138621 运行的最新结果:
python { "all": { "acc": 0.6543450273126857, "acc_stderr": 0.03191864171781636, "acc_norm": 0.6545137141283983, "acc_norm_stderr": 0.03257628315307556, "mc1": 0.39167686658506734, "mc1_stderr": 0.01708779588176963, "mc2": 0.558695592929387, "mc2_stderr": 0.015276769304708891 }, "harness|arc:challenge|25": { "acc": 0.6390784982935154, "acc_stderr": 0.014034761386175456, "acc_norm": 0.6791808873720137, "acc_norm_stderr": 0.013640943091946533 }, "harness|hellaswag|10": { "acc": 0.6694881497709619, "acc_stderr": 0.004694360968929403, "acc_norm": 0.8598884684325832, "acc_norm_stderr": 0.003463933286063885 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.33, "acc_stderr": 0.04725815626252605, "acc_norm": 0.33, "acc_norm_stderr": 0.04725815626252605 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- data_files:
- split: 2023_12_12T03_53_21.138621
- **/details_harness|arc:challenge|25_2023-12-12T03-53-21.138621.parquet
- split: latest
- **/details_harness|arc:challenge|25_2023-12-12T03-53-21.138621.parquet
- split: 2023_12_12T03_53_21.138621
- data_files:
-
config_name: harness_gsm8k_5
- data_files:
- split: 2023_12_12T03_53_21.138621
- **/details_harness|gsm8k|5_2023-12-12T03-53-21.138621.parquet
- split: latest
- **/details_harness|gsm8k|5_2023-12-12T03-53-21.138621.parquet
- split: 2023_12_12T03_53_21.138621
- data_files:
-
config_name: harness_hellaswag_10
- data_files:
- split: 2023_12_12T03_53_21.138621
- **/details_harness|hellaswag|10_2023-12-12T03-53-21.138621.parquet
- split: latest
- **/details_harness|hellaswag|10_2023-12-12T03-53-21.138621.parquet
- split: 2023_12_12T03_53_21.138621
- data_files:
-
config_name: harness_hendrycksTest_5
- data_files:
- split: 2023_12_12T03_53_21.138621
- **/details_harness|hendrycksTest-abstract_algebra|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-anatomy|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-astronomy|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-college_biology|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-college_physics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-computer_security|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-econometrics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-global_facts|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-human_aging|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-international_law|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-management|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-marketing|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-nutrition|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-philosophy|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-prehistory|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-professional_accounting|5_2023-12-12T03-53-21.138621.parquet
- **/details_harness|hendrycksTest-professional_law|5_2023-12-12T03-53-21.138621
- split: 2023_12_12T03_53_21.138621
- data_files:



