Constitutional Midtraining: Content Presence Drives Alignment Gains
Quick summary
arXiv:2607.26654v2 Announce Type: cross Abstract: Post-training alignment is often shallow, eroding under fine-tuning. It remains untested as to whether constitutional midtraining interventions can produce durable alignment when cleanly isolated from post-training. We build a 394M-token constitutional corpus from Anthropic's Constitution and apply constitutional midtraining at 120B scale, where principled, values-based content is inserted into midtraining. A 2x2 design (curriculum ordering x deliberative reasoning) was used to produce four constitutionally midtrained conditions, plus a control
Key takeaways
- arXiv:2607.26654v2 Announce Type: cross Abstract: Post-training alignment is often shallow, eroding under fine-tuning.
- It remains untested as to whether constitutional midtraining interventions can produce durable alignment when cleanly isolated from post-training.
- We build a 394M-token constitutional corpus from Anthropic's Constitution and apply constitutional midtraining at 120B scale, where principled, values-based content is inserted into midtraining.
Why it matters
This model development creates a new option for users and a new testing obligation for developers. A fixed evaluation set comparing quality, cost and failure behavior is more useful than launch claims.