AI RESEARCH

Interactive Critique-Revision Training for Reliable Structured LLM Generation

arXiv CS.AI

ArXi:2605.08327v1 Announce Type: cross In structured decision-making workflows such as form filling, compliance checking, and maintenance reporting, LLM outputs must be locally correct, globally consistent, and auditable against task-specific rules. Existing refinement methods often rely on heuristic debate, self-play, or LLM-generated supervision, creating a second-order assurance problem. We propose DPA-GRPO (Dual Paired-Action Group-Relative Policy Optimization), a paired-action