Safety Reconstructed: Generative Modeling via Masked Diffusion Builds Strong Safety Guardrails
This work proposes a masked diffusion language model, fine-tuning LLaDA-8B-Instruct with a class-conditional reconstruction objective using LoRA and requiring no architectural changes beyond the base model, which leads on average rank against discriminative baselines trained on stronger backbones across seven held-out...
Gert Lek, Abele Malan, Chao-Yi Zhu et al.
· 0 citations