1
One Token per Multimodal Evidence: Latent Memory for Resource-Constrained QA
多模态问答模型资源受限新方案,仅用单token压缩证据降低内存开销。
arXiv:2606.10572v1 Announce Type: new Abstract: External memory effectively grounds large language models (LLMs) and vision-language models (VLMs)-bas…