A False Average: Pooled CoT-Monitor Accuracy Conceals a Reasoning-Dependent Fragility

arXiv:2608.00583v3 Announce Type: replace-cross Abstract: Chain-of-thought (CoT) monitors are reported by their aggregate accuracy on a pool of reward hacks. We show that this number is a false average. On Terminal Wrench, about 77% of hacks are…

Read the full story at arXiv cs.AI ↗