跳到正文
原文
ArtificialAnlys· @ArtificialAnlys · X·· 4 小时前AI 评分32

AA-Omniscience v1.1 新增难度匹配公开测试集

AI 导读

AA-Omniscience v1.1 在 @huggingface 上新增了一个公开测试集,其难度经过校准,与私有测试集相匹配。发布时,公开集与私有集在 AA-Omniscience Net Score 上的相关性为 0.99,分数中位数差异为 0.0 分。准确率和模型幻觉率的相关性也均为 0.99,中位数差异分别为 -0.07 和 -0.29 个百分点。

正文 · 原文

AA-Omniscience v1.1 has a new public set on @huggingface calibrated to match the difficulty of the private set. At release, public and private scores have a 0.99 correlation for AA-Omniscience Net Score, with a median difference of 0.0 points. Accuracy and hallucination rate also have 0.99 correlations, with median differences of -0.07 and -0.29 percentage points, respectively

来源:ArtificialAnlys · x.com