Salvato in:
Dettagli Bibliografici
Autori principali: Kong, Injin, Lee, Hyoungjoon, Jo, Yohan
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2601.14758
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910269902422016
author Kong, Injin
Lee, Hyoungjoon
Jo, Yohan
author_facet Kong, Injin
Lee, Hyoungjoon
Jo, Yohan
contents Post-training pretrained autoregressive models (ARMs) into masked diffusion models (MDMs) has emerged as a cost-effective way to overcome the limitations of sequential generation. Yet it remains unclear whether post-trained MDMs acquire genuinely new computational mechanisms or merely re-express autoregressive computation in a non-autoregressive form. Through a comparative circuit analysis of ARMs and their MDM counterparts post-trained from the same backbones, we uncover two complementary axes of reorganization. Structurally, the shift is task-dependent: MDMs preserve autoregressive circuitry on locally causal tasks but abandon inherited pathways and front-load computation into early layers on global tasks. Semantically, the shift is consistent across regimes: sharp, localized specialization in ARMs gives way to distributed integration in MDMs. Together, these findings show that diffusion post-training is not a surface-level change in the generation procedure but a reorganization of internal computation whose depth depends on the task.
format Preprint
id arxiv_https___arxiv_org_abs_2601_14758
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Mechanism Shift During Post-training from Autoregressive to Masked Diffusion Language Models
Kong, Injin
Lee, Hyoungjoon
Jo, Yohan
Machine Learning
Artificial Intelligence
Computation and Language
Post-training pretrained autoregressive models (ARMs) into masked diffusion models (MDMs) has emerged as a cost-effective way to overcome the limitations of sequential generation. Yet it remains unclear whether post-trained MDMs acquire genuinely new computational mechanisms or merely re-express autoregressive computation in a non-autoregressive form. Through a comparative circuit analysis of ARMs and their MDM counterparts post-trained from the same backbones, we uncover two complementary axes of reorganization. Structurally, the shift is task-dependent: MDMs preserve autoregressive circuitry on locally causal tasks but abandon inherited pathways and front-load computation into early layers on global tasks. Semantically, the shift is consistent across regimes: sharp, localized specialization in ARMs gives way to distributed integration in MDMs. Together, these findings show that diffusion post-training is not a surface-level change in the generation procedure but a reorganization of internal computation whose depth depends on the task.
title Mechanism Shift During Post-training from Autoregressive to Masked Diffusion Language Models
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2601.14758