Evidence · DiggingBeagle record
OpenAI: CoT monitorability changes through pretraining and reinforcement learning
Original research on no-chain-of-thought capability and reasoning-trace controllability, distinct from observed full trajectory monitor evasion.
- Published
- Nov 10, 2026
- Source role
- primary disclosure
Evidence record
Original research on no-chain-of-thought capability and reasoning-trace controllability, distinct from observed full trajectory monitor evasion.
Claim-level citations (2)
- supportsModel reasoning may become easier to manipulate without necessarily becoming better at evading a full action monitor: OpenAI studied no-chain-of-thought task ability and the capacity to steer emitted reasoning through model training. Some pretraining and architectural effects improved no-CoT performance; outcome-reward reinforcement learning and generated teacher data increased control over visible reasoning text. The paper tested matched-capability models and found a divergence: higher reasoning controllability did not produce greater full-trajectory monitoring evasion on a selected math side-task and was associated with less success at prompted sandbagging in that evaluation.
OpenAI Alignment November 10 report, RL and synthetic data experiments, full-trajectory evasion comparison
- supportsModel reasoning may become easier to manipulate without necessarily becoming better at evading a full action monitor: These metrics are proxies for monitorability and are not a verified rate of hidden real-world agent attacks. A model having the ability to alter verbal reasoning does not establish a motive, unauthorized effect or successful concealment. The security boundary is between optionally emitted model reasoning and independent trusted records of actions, permissions and effects.
Primary disclosure: results, scope and limitations