AINeutralarXiv – CS AI · 14h ago6/10
🧠
The Confidence Shortcut: A Reasoning Failure Mode of Masked Diffusion Models
Researchers identify a critical failure mode in masked diffusion language models where confidence-based decoding strategies cause reasoning errors on complex tasks. The study demonstrates that confidence-aligned training amplifies these failures by an order of magnitude, while random masking preserves robust reasoning capabilities across five reasoning tasks.