DeepSeek V4 Keeps Student Rollouts In-Distribution Unlike Nvidia Ultra

7/ In other words: V4 says sampled-token is too weak to carry the merge, while Ultra says full-vocab is worse than sampled-token. With high overlap the dense loss is safe and useful; with low overlap it's harmful.

4/ Nvidia Ultra doesn't. Its teachers get their skills from SFT on data generated by external models (V4-Pro, gpt-oss, GLM) the student never saw. That pushes each teacher's distribution away from the student, so student rollouts can be out-of-distribution for the teacher.

5/ This is why Ultra needs a "warmup" SFT to pull the student onto teacher support before distilling, while V4 doesn't.

6/ It also decides the loss. V4 matches the full teacher distribution at every token. In their report, Ultra tried that, top-k and full vocab, and found it worse than scoring only the sampled token, especially on agentic tasks (It amplifies the noise).

8/ You can see it in their RL too. V4 drops student RL entirely because shared lineage already keeps rollouts on-support. Ultra keeps unified RLVR partly so the student samples stronger rollouts that land closer to teacher support.