MuPHI: Learning Implicit Multimodal Harm Reasoning via Semantically Grounded Reward Optimization 文章

arXiv:2605.29951v1 Announce Type: cross Abstract: Understanding how harm emerges from interaction between otherwise benign image-text pairs requires intent-aware cross-modal reasoning beyond surface-level features. Existing vision-language models (VLMs) excel at literal reasoning over perceptual cues but often fail to derive harmful semantics that rely on implicit, context-dependent reasoning. To evaluate VLMs on compositional harm detection and reasoning, we introduce Multimodal Pragmatic Harm Interpretation (MuPHI), a dataset containing image-text pairs where harm is encoded in subtle multimodal cues. MuPHI spans diverse harm categories and includes annotated harm rationales for assessing VLM reasoning chains. To improve both detection and reasoning in VLMs, we propose MuPHIRM, a reasoning-augmented training framework which learns joint semantics by optimizing multi-perspective rewards.