AI-generated music detectors can appear robust on standard benchmark splits, yet their deployments require transfer to generator sources absent during training. We study this problem with source-restricted evaluation on \textsc{MoM-open}, an open reconstruction of MoM-CLAM that replaces the non-redistributable real corpus with FMA and MTG-Jamendo while preserving the fake-generator protocol. To isolate the role of representation, we introduce \textsc{CoMoE}, a compact fixed classifier for comparing heterogeneous audio token spaces while keeping the downstream architecture and training recipe unchanged. Experiments show that standard and real-source-restricted splits are nearly saturated, whereas fake-source restriction exposes large differences between token spaces: X-Codec tokens are strongest when training on Udio alone, while MERT-derived tokens are stronger when training on Suno-v3.5 alone. These results suggest that codec-style discrete token spaces should be treated as a primary experimental axis under generator shift in AI-generated music detection. Our code and data are available at https://github.com/MAAP-LAB/CoMoE.

翻译：AI生成音乐检测器在标准基准划分上可能表现鲁棒，但其部署需要迁移至训练期间未出现的生成器来源。我们通过源受限评估研究该问题，基于MoM-open，这是MoM-CLAM的开源重构，用FMA和MTG-Jamendo替代不可再分发的真实语料库，同时保留虚假生成器协议。为隔离表示的作用，我们引入CoMoE，一个紧凑的固定分类器，用于在保持下游架构和训练方案不变的情况下比较异构音频令牌空间。实验表明，标准和真实源受限划分几乎饱和，而虚假源受限则揭示了令牌空间之间的巨大差异：在仅用Udio训练时，X-Codec令牌表现最强，而仅用Suno-v3.5训练时，MERT导出令牌表现更强。这些结果表明，在AI生成音乐检测的生成器偏移下，编解码器风格的离散令牌空间应被视为主要实验轴线。我们的代码和数据可在https://github.com/MAAP-LAB/CoMoE获取。