2 papers
cs.LG2026
Do Understanding and Generation Fight? A Diagnostic Study of DPO for Unified Multimodal Models
Abinav Rao, Sujan Rachuri
Unified multimodal models share a language model backbone for both understanding and generating images. Can DPO align both capabilities simultaneously? We present the first systema…
cs.CL2026
Correct Chains, Wrong Answers: Dissociating Reasoning from Output in LLM Logic
Abinav Rao, Sujan Rachuri, Nikhil Vemuri
LLMs can execute every step of chain-of-thought reasoning correctly and still produce wrong final answers. We introduce the Novel Operator Test, a benchmark that separates operator…