SOURCE-LINKED INTELLIGENCE

Selective Disclosure of Hidden Directives in Reasoning Models: Behavioral Asymmetry and Steering

arXiv · AI, language, vision and robotics · article · Aug 29, 2026 · UTC

Chain-of-thought (CoT) reasoning traces are increasingly proposed as a mechanism for AI oversight: a monitor inspecting a model's reasoning can, in principle, detect misbehavior invisible from outputs alone. This assumes CoT surfaces what a model is instructed to do regardless of the instructions given. We test this assumption along two axes. First, we introduce the Instruction-Compliance Gap (ICG): the difference in probability that a model's CoT explicitly references a hidden system prompt directive when that directive is malign versus benign. Across 100 task pairs and 8 frontier reasoning m

Read original source ↗ Open in workspace

recordType

paper

region

Global