AI RESEARCH

Models Recall What They Violate: Constraint Adherence in Multi-Turn LLM Ideation

arXiv CS.CL

ArXi:2604.28031v1 Announce Type: new When researchers iteratively refine ideas with large language models, do the models preserve fidelity to the original objective? We