2 papers
cs.LG2026
Layer Collapse in Diffusion Language Models
Alexander Conzelmann, Albert Catalan-Tatjer, Shiwei Liu
Diffusion language models (DLMs) have recently emerged as competitive alternatives to autoregressive (AR) language models, yet differences in their activation dynamics remain poorl…
cs.LG2026
Training Dynamics Impact Post-Training Quantization Robustness
Albert Catalan-Tatjer, Niccolò Ajroldi, Jonas Geiping
While post-training quantization is widely adopted for efficient deployment of large language models, the mechanisms underlying quantization robustness remain unclear. We conduct a…