{"id":"2602.03134","title":"SwiftVLM: Efficient Vision-Language Model Inference via Cross-Layer Token Bypass","url":"https://codeofpaper.com/paper/2602.03134","arxiv_url":"https://arxiv.org/abs/2602.03134","published":"2026-02-03","code":"none","best":null,"most_starred_community":null,"implementations":[],"possible_matches":0}