HaloProbe introduces a Bayesian method to estimate token-level hallucination probabilities in VLMs by factorizing external and internal signals, enabling more effective mitigation than intervention-based techniques while preserving utility.
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
-
HaloProbe: Bayesian Detection and Mitigation of Object Hallucinations in Vision-Language Models
HaloProbe introduces a Bayesian method to estimate token-level hallucination probabilities in VLMs by factorizing external and internal signals, enabling more effective mitigation than intervention-based techniques while preserving utility.