Good performance can be maintained with only the top item in DiffusionGemma, supporting high monitorability

We strengthen these results by showing that this performance degradation is largely a sampler artifact and good performance can be maintained with only the top item, supporting the case for high monitorability.

https://www.alignmentforum.org/posts/QBuJ3suRZxrrxSTtv/does-diffusiongemma-do-latent-reasoning