Proposed
Public long-trace jailbreak suites
Do not let ‘we refuse in the system card’ stand in for ‘we refuse at step 80’.
$24M experimental capital · 2 private / 0 institute
Who should own it
AISI / METR
Alliances / evaluators
How quickly it can land
Weeks
A dedicated squad can land it inside two months.
Expedited implementation
4 weeks
30 calendar days with a crash team
Normal implementation
4 months
120 calendar days as a planned program
Risk this mitigates
15Given that research shows longer chain-of-thought dilutes refusal and lifts jailbreak success toward 80% across major models, there is a possibility of safety training that holds in short chats failing the moment a user or agent reasons at length resulting in every other harmful capability on this register becoming available through a conversational side door.
Residual composite 15 · still above the threshold
Effect if implemented
Applied to every failure scenario on that risk, then re-ranked. Axes are clamped at 1.
- Likelihood
- −0
- Consequence
- −0
- Urgency
- −1