open-llm-leaderboard-old/details_AA051615__L0223
收藏数据集概述
该数据集是在模型 AA051615/L0223 的评估运行期间自动创建的,用于 Open LLM Leaderboard。数据集包含 63 个配置,每个配置对应一个评估任务。
数据集结构
- 创建来源:数据集从 1 次运行中创建。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。
- 分割:每个配置包含多个分割,其中 "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 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_AA051615__L0223", "harness_winogrande_5", split="train")
最新结果
以下是 2024-02-24T21:40:39.442085 运行的最新结果:
python { "all": { "acc": 0.8158382341774466, "acc_stderr": 0.02573910232891644, "acc_norm": 0.8235047224734193, "acc_norm_stderr": 0.026168932766623194, "mc1": 0.3635250917992656, "mc1_stderr": 0.016838862883965824, "mc2": 0.535935780967817, "mc2_stderr": 0.015547988224447432 }, "harness|arc:challenge|25": { "acc": 0.6407849829351536, "acc_stderr": 0.014020224155839159, "acc_norm": 0.6791808873720137, "acc_norm_stderr": 0.013640943091946528 }, "harness|hellaswag|10": { "acc": 0.6333399721171081, "acc_stderr": 0.004809077205343492, "acc_norm": 0.8299143596893049, "acc_norm_stderr": 0.003749401775087307 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.6, "acc_stderr": 0.04923659639173309, "acc_norm": 0.6, "acc_norm_stderr": 0.04923659639173309 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- split: 2024_02_24T21_40_39.442085
**/details_harness|arc:challenge|25_2024-02-24T21-40-39.442085.parquet
- split: latest
**/details_harness|arc:challenge|25_2024-02-24T21-40-39.442085.parquet
- split: 2024_02_24T21_40_39.442085
-
config_name: harness_gsm8k_5
- split: 2024_02_24T21_40_39.442085
**/details_harness|gsm8k|5_2024-02-24T21-40-39.442085.parquet
- split: latest
**/details_harness|gsm8k|5_2024-02-24T21-40-39.442085.parquet
- split: 2024_02_24T21_40_39.442085
-
config_name: harness_hellaswag_10
- split: 2024_02_24T21_40_39.442085
**/details_harness|hellaswag|10_2024-02-24T21-40-39.442085.parquet
- split: latest
**/details_harness|hellaswag|10_2024-02-24T21-40-39.442085.parquet
- split: 2024_02_24T21_40_39.442085
-
config_name: harness_hendrycksTest_5
- split: 2024_02_24T21_40_39.442085
**/details_harness|hendrycksTest-abstract_algebra|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-anatomy|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-astronomy|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-college_biology|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-college_physics|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-computer_security|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-econometrics|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-global_facts|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-02-24T21-40-39.442085.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-02-24T21-40-39.442085.parquet
其他文件路径...
- split: 2024_02_24T21_40_39.442085
}



