arXiv:2610.15855v1 Announce Type: new Abstract: We study how to detect and mitigate visual attention sinks in vision-language models (VLMs).
← Back to all developments · Signals are a development reported by more than one source. See the methodology.