open-llm-leaderboard-old/details_tourist800__Marcoro14-7B-slerp
收藏数据集概述
该数据集是在对模型 tourist800/Marcoro14-7B-slerp 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。
- 数据集来自 1 次运行,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。
- 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 2024-01-28T18:32:24.206889 运行的最新结果:
python { "all": { "acc": 0.6111388777526852, "acc_stderr": 0.03287383799644916, "acc_norm": 0.6159662135212005, "acc_norm_stderr": 0.03353760847602086, "mc1": 0.36107711138310894, "mc1_stderr": 0.016814312844836882, "mc2": 0.5207873423930568, "mc2_stderr": 0.0153889376471881 }, "harness|arc:challenge|25": { "acc": 0.5861774744027304, "acc_stderr": 0.014392730009221004, "acc_norm": 0.6339590443686007, "acc_norm_stderr": 0.014077223108470137 }, "harness|hellaswag|10": { "acc": 0.6421031666998606, "acc_stderr": 0.004784018497679814, "acc_norm": 0.8376817367058355, "acc_norm_stderr": 0.0036798891253998155 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.26, "acc_stderr": 0.0440844002276808, "acc_norm": 0.26, "acc_norm_stderr": 0.0440844002276808 }, # 其他任务的结果... }
配置详情
-
harness_arc_challenge_25
- 分割:2024_01_28T18_32_24.206889
- 路径:
**/details_harness|arc:challenge|25_2024-01-28T18-32-24.206889.parquet - 分割:latest
- 路径:
**/details_harness|arc:challenge|25_2024-01-28T18-32-24.206889.parquet
-
harness_gsm8k_5
- 分割:2024_01_28T18_32_24.206889
- 路径:
**/details_harness|gsm8k|5_2024-01-28T18-32-24.206889.parquet - 分割:latest
- 路径:
**/details_harness|gsm8k|5_2024-01-28T18-32-24.206889.parquet
-
harness_hellaswag_10
- 分割:2024_01_28T18_32_24.206889
- 路径:
**/details_harness|hellaswag|10_2024-01-28T18-32-24.206889.parquet - 分割:latest
- 路径:
**/details_harness|hellaswag|10_2024-01-28T18-32-24.206889.parquet
-
harness_hendrycksTest_5
- 分割:2024_01_28T18_32_24.206889
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-anatomy|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-astronomy|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-college_biology|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-college_physics|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-computer_security|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-econometrics|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-global_facts|5_2024-01-28T18-32-24.206889.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-01-28T18-32-24.206889.parquet
其他路径...
}
数据集加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_tourist800__Marcoro14-7B-slerp", "harness_winogrande_5", split="train")



