open-llm-leaderboard-old/details_ValiantLabs__Esper-70b
收藏数据集概述
数据集摘要
该数据集是在对模型 ValiantLabs/Esper-70b 进行评估运行期间自动创建的,用于 Open LLM Leaderboard。
数据集组成
- 数据集包含 63 个配置,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" 分割始终指向最新结果。
- 额外的 "results" 配置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。
最新结果
以下是 2024-03-15T10:12:58.393533 运行的最新结果:
python { "all": { "acc": 0.5598497316876694, "acc_stderr": 0.034162766283638, "acc_norm": 0.5628921495292818, "acc_norm_stderr": 0.03486354447067593, "mc1": 0.31701346389228885, "mc1_stderr": 0.016289203374403392, "mc2": 0.4598011127457919, "mc2_stderr": 0.015071082364048416 }, "harness|arc:challenge|25": { "acc": 0.5196245733788396, "acc_stderr": 0.014600132075947094, "acc_norm": 0.5648464163822525, "acc_norm_stderr": 0.014487986197186047 }, "harness|hellaswag|10": { "acc": 0.582055367456682, "acc_stderr": 0.0049221295689195815, "acc_norm": 0.7772356104361681, "acc_norm_stderr": 0.004152510556342204 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.37, "acc_stderr": 0.04852365870939098, "acc_norm": 0.37, "acc_norm_stderr": 0.04852365870939098 }, # 其他任务的结果... }
配置详情
-
config_name: harness_arc_challenge_25
- split: 2024_03_15T10_12_58.393533
- path: **/details_harness|arc:challenge|25_2024-03-15T10-12-58.393533.parquet
- split: latest
- path: **/details_harness|arc:challenge|25_2024-03-15T10-12-58.393533.parquet
- split: 2024_03_15T10_12_58.393533
-
config_name: harness_gsm8k_5
- split: 2024_03_15T10_12_58.393533
- path: **/details_harness|gsm8k|5_2024-03-15T10-12-58.393533.parquet
- split: latest
- path: **/details_harness|gsm8k|5_2024-03-15T10-12-58.393533.parquet
- split: 2024_03_15T10_12_58.393533
-
config_name: harness_hellaswag_10
- split: 2024_03_15T10_12_58.393533
- path: **/details_harness|hellaswag|10_2024-03-15T10-12-58.393533.parquet
- split: latest
- path: **/details_harness|hellaswag|10_2024-03-15T10-12-58.393533.parquet
- split: 2024_03_15T10_12_58.393533
-
config_name: harness_hendrycksTest_5
- split: 2024_03_15T10_12_58.393533
- path:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-03-15T10-12-58.393533.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2024-03-15T10-12-58.393533.parquet
其他路径...
- path:
- split: 2024_03_15T10_12_58.393533



