Tag
SLIM-RL introduces a risk-budgeted random-masking reinforcement learning method for diffusion LLMs that avoids trajectory slicing, achieving state-of-the-art results on math and code benchmarks with significantly fewer training samples.
This paper identifies a failure mode in masked diffusion language models where confidence-based decoding leads to high-confidence errors on complex reasoning tasks, and shows that confidence-aligned training exacerbates this issue while random masking preserves reasoning performance.