open-llm-leaderboard-old/details_RaduGabriel__SirUkrainian
收藏数据集概述
该数据集是在对模型 RaduGabriel/SirUkrainian 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" 分割始终指向最新的结果。
- 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_RaduGabriel__SirUkrainian", "harness_winogrande_5", split="train")
最新结果
以下是 2024-02-15T16:52:40.545415 运行的最新结果:
python { "all": { "acc": 0.6341358887316704, "acc_stderr": 0.03253233597406997, "acc_norm": 0.6358767386408157, "acc_norm_stderr": 0.03319318434906525, "mc1": 0.5324357405140759, "mc1_stderr": 0.017466632149577613, "mc2": 0.6873616277110236, "mc2_stderr": 0.014862338695256647 }, "harness|arc:challenge|25": { "acc": 0.6348122866894198, "acc_stderr": 0.014070265519268802, "acc_norm": 0.6732081911262798, "acc_norm_stderr": 0.013706665975587328 }, "harness|hellaswag|10": { "acc": 0.676956781517626, "acc_stderr": 0.004666833452796188, "acc_norm": 0.8554072893845848, "acc_norm_stderr": 0.0035097096477918377 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.28, "acc_stderr": 0.045126085985421276, "acc_norm": 0.28, "acc_norm_stderr": 0.045126085985421276 }, ... }
配置详情
-
config_name: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_02_15T16_52_40.545415
- 路径:
**/details_harness|arc:challenge|25_2024-02-15T16-52-40.545415.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|arc:challenge|25_2024-02-15T16-52-40.545415.parquet
- 路径:
- 分割: 2024_02_15T16_52_40.545415
- 数据文件:
-
config_name: harness_gsm8k_5
- 数据文件:
- 分割: 2024_02_15T16_52_40.545415
- 路径:
**/details_harness|gsm8k|5_2024-02-15T16-52-40.545415.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|gsm8k|5_2024-02-15T16-52-40.545415.parquet
- 路径:
- 分割: 2024_02_15T16_52_40.545415
- 数据文件:
-
config_name: harness_hellaswag_10
- 数据文件:
- 分割: 2024_02_15T16_52_40.545415
- 路径:
**/details_harness|hellaswag|10_2024-02-15T16-52-40.545415.parquet
- 路径:
- 分割: latest
- 路径:
**/details_harness|hellaswag|10_2024-02-15T16-52-40.545415.parquet
- 路径:
- 分割: 2024_02_15T16_52_40.545415
- 数据文件:
-
config_name: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_02_15T16_52_40.545415
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-anatomy|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-astronomy|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-college_biology|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-college_physics|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-computer_security|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-econometrics|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-global_facts|5_2024-02-15T16-52-40.545415.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-02-15T16-52-40.545415.parquet- ... }
- 路径:
- 分割: 2024_02_15T16_52_40.545415
- 数据文件:



