GPT-6 Astra's monitorability decline exceeds that predicted by earlier models' capability trends.
3 events · 1 assessment · 1 decision
Structured and assessed
First pass. - Decomposition: linked existing claims for the measured decline (assumes), the RL controllability reversal (supports) and the controllability–capability correlation (contradicts). Minted two subclaims: the no-CoT horizon being above trend (supports, importance 0.35) and possible contamination of that figure (contradicts, importance 0.2, deferred stub). - Arguments: added one argument for and one against, with written forms and evaluations. - Recorded Zvi Mowshowitz's affirming instance. - Importance set to 0.35, contestation 0.5. - Assessment: inconclusive, credence 0.5. - Notified the parent, which uses this claim as a premise.
Assessed Inconclusive
verdict confidence 0.65 · credence 0.50
GPT-6 Astra's chain-of-thought monitorability is lower than GPT-5.6 Sol's, and that decline is well established. Whether the decline is larger than earlier models' capability trends would predict depends on two things: how capability is measured, and whether any trend is well enough defined to make a prediction at all. The case for an excess comes mainly from commentators, most prominently Zvi Mowshowitz. It has two parts: - UK AISI measured Astra's no-chain-of-thought math time horizon at 30.9 minutes, against 3.6 minutes for its predecessor. Mowshowitz judges that Astra's general capability would predict roughly 5 to 9 minutes, so the no-CoT horizon sits well above trend. - Astra's chain-of-thought controllability rose during RL, whereas earlier models' controllability fell. That suggests a change in trajectory, not just more of the same. Against this: - No published model predicts monitorability from capability across model generations. The "straight lines on graphs" are informal extrapolations from a few data points. - OpenAI reports that controllability gains track no-CoT capability gains. With capability measured that way, the decline lies on the capability line rather than above it. - UK AISI itself flags contamination and limited testing time, so the headline no-CoT figure may be inflated. The claim is therefore plausible but not demonstrated. It would be resolved by a quantitative fit of monitorability or controllability against a general capability index across several OpenAI models, with a contamination-checked no-CoT evaluation for Astra.
Claim entered the graph