Alignment Midtraining Cracks Under Pressure
Key takeaway
TL;DRWe stress-test alignment midtraining (AMT) across model and token budget scales.
Why it's on PDOOM
PDOOM selected this story for its alignment & control signals: Alignment.
AlignmentFrom LessWrong AI
TL;DRWe stress-test alignment midtraining (AMT) across model and token budget scales. Our results suggest that midtraining cannot tackle the hard problems of AI alignment—namely distributional shift and reward underspecification in the presence of imperfect data.For instance, we test whether midtrained motivations are robust to finetuning which elicits competing motivations. In our setting, 190M tokens of midtrained motivations are overpowered by a relatively tiny amount (~50K tokens) of competi
By J Bostock