AI RESEARCH
Masked by Consensus: Disentangling Privileged Knowledge in LLM Correctness
arXiv CS.CL
•
ArXi:2604.12373v1 Announce Type: new