IKOD reduces hallucination in vision-language models by merging KV states to derive image-focused shorter-sequence logits and combining them with normal decoding, without training.
E.; Stoica, I.; and Xing, E
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
IKOD: Mitigating Visual Attention Degradation in Large Vision-Language Models
IKOD reduces hallucination in vision-language models by merging KV states to derive image-focused shorter-sequence logits and combining them with normal decoding, without training.