AI RESEARCH

Masked by Consensus: Disentangling Privileged Knowledge in LLM Correctness

arXiv CS.CL

ArXi:2604.12373v1 Announce Type: new