arXivPedro Ferreira, Wilker Aziz, Ivan TitovWed, Aug 5, 2026, 7:55 AM PDT
score 17.2
Hidden chain-of-thought reasoning can still be monitored for bias
Original: Does Out-of-Sight Equal Out-of-Mind in CoT Monitorability?
Source: arxiv.org ↗
Writing ELI5 summary…