Unveiling the Response of Large Vision-Language Models to Visually Absent Tokens

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.