five

strategy-scope/res_exp_best-config-165-gpt4.1mini_20260407_111647-20260407_123541

收藏
Hugging Face2026-04-07 更新2026-04-12 收录
下载链接:
https://hf-mirror.com/datasets/strategy-scope/res_exp_best-config-165-gpt4.1mini_20260407_111647-20260407_123541
下载链接
链接失效反馈
官方服务:
资源简介:
--- tags: - strategy-scope - CREATE - evaluation --- # res_exp_best-config-165-gpt4.1mini_20260407_111647-20260407_123541 Evaluation results for `strategy-scope/exp_best-config-165-gpt4.1mini_20260407_111647`. ## Aggregate Statistics | Metric | Value | |--------|-------| | Instances | 165 | | Avg paths/instance | 18.1 | | Avg valid/instance | 17.4 | | Avg valid & factual/instance | 5.7 | | Avg factuality | 0.6833 | | Avg strength | 2.5744 | | Avg pairwise distance (ft=0.0) | 0.7504 | | Avg pairwise distance (ft=1.0) | 0.6491 | | Avg utility (ft=0.0) | 19.1919 | | Avg utility (ft=1.0) | 8.7548 | ## Parameters - **Eval model:** gpt-4o-mini - **Patience:** 0.9 - **Total eval calls:** 5962 - **Timestamp:** 20260407_123541
提供机构:
strategy-scope
二维码
社区交流群
二维码
科研交流群
商业服务