Immuno-VLM generates semantic antibodies via LLMs to bound VLM decision spaces for open-world trustworthiness and reports new SOTA results on ImageNet-1K plus four OOD benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Immuno-VLM: Immunizing Large Vision-Language Models via Generative Semantic Antibodies for Open-World Trustworthiness
Immuno-VLM generates semantic antibodies via LLMs to bound VLM decision spaces for open-world trustworthiness and reports new SOTA results on ImageNet-1K plus four OOD benchmarks.