The argument from missed conditional misalignment
The argument
Standard batteries ask context-free questions with no grader or training-like cue in view. Misalignment that is conditional on such cues goes unregistered: reward-hacking-trained models misbehave where a grader is salient while scoring near zero on the batteries, and interventions that zero standard scores leave conditional misalignment on training-like prompts. A measure that misses real misalignment in these cases understates it there, so the error is not uniformly toward overestimation.
Does it work?
- Does it follow?
- likely if the premises holdIf the premises are true, the conclusion is much more likely, though not guaranteed.constitution →
- Do the premises hold?
- 2 supported
Granting its premises, the standard batteries understate misalignment in at least some fine-tuned models, which tells against a general claim of overestimation, though not against overestimation for other models or items. Its strongest premise, that batteries miss grader-salient misalignment, is supported but rests largely on one run. The premise that conditional misalignment survives interventions is not yet assessed.
Premises
- SupportedThe evidence favors the claim, but the examination is incomplete or the evidence is indirect.constitution →Standard emergent-misalignment evaluations miss misalignment in reward-hacking-trained models that appears only where a reward, grader or evaluation is salient.
- SupportedThe evidence favors the claim, but the examination is incomplete or the evidence is indirect.constitution →Interventions that suppress emergent misalignment on standard evaluations leave conditional misalignment on training-like prompts.
Where it is made
Passages where sources make, criticize or mention this argument. They record where it appears, not how strong it is, and none has precedence over another.
No source is recorded as making it.