UnMaskFork: Test-Time Scaling for Masked Diffusion via Deterministic Action Branching

（＊日本語は英文の後に） Can test-time scaling work for diffusion language models? In our ICML 2026 paper “UnMaskFork,” we show that having multiple masked diffusion language models collaborate on a single answer improves performance on coding and math tasks. Test-time scaling is an actively researched technique that boosts LLM performance by using inference-time compute, for example, by having a model think longer or repeatedly refine its answers.