Preprint
Aug 2026
Wiener Representation Filtering for VLM Hallucination Suppression
The generality of this approach is demonstrated on the TempCompass video understanding benchmark and on discrete diffusion language models for grounded dialogue, showing that representation filtering reduces hallucinations even in temporal video reasoning and multi-step, sequence-wide denoising settings.
Ameen Ali, Tamim Zoabi, Lidor Brami et al.
· 0 citations