Beyond Hard Masks: Progressive Token Evolution for Diffusion Language Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911378570215424 |
|---|---|
| author | Zhong, Linhao Wu, Linyu Fang, Bozhen Feng, Tianjian Jing, Chenchen Wang, Wen Zhang, Jiaheng Chen, Hao Shen, Chunhua |
| author_facet | Zhong, Linhao Wu, Linyu Fang, Bozhen Feng, Tianjian Jing, Chenchen Wang, Wen Zhang, Jiaheng Chen, Hao Shen, Chunhua |
| contents | Diffusion Language Models (DLMs) offer a promising alternative for language modeling by enabling parallel decoding through iterative refinement. However, most DLMs rely on hard binary masking and discrete token assignments, which hinder the revision of early decisions and underutilize intermediate probabilistic representations. In this paper, we propose EvoToken-DLM, a novel diffusion-based language modeling approach that replaces hard binary masks with evolving soft token distributions. EvoToken-DLM enables a progressive transition from masked states to discrete outputs, supporting revisable decoding. To effectively support this evolution, we introduce continuous trajectory supervision, which aligns training objectives with iterative probabilistic updates. Extensive experiments across multiple benchmarks show that EvoToken-DLM consistently achieves superior performance, outperforming strong diffusion-based and masked DLM baselines. Project webpage: https://aim-uofa.github.io/EvoTokenDLM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_07351 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Beyond Hard Masks: Progressive Token Evolution for Diffusion Language Models Zhong, Linhao Wu, Linyu Fang, Bozhen Feng, Tianjian Jing, Chenchen Wang, Wen Zhang, Jiaheng Chen, Hao Shen, Chunhua Computation and Language Artificial Intelligence Diffusion Language Models (DLMs) offer a promising alternative for language modeling by enabling parallel decoding through iterative refinement. However, most DLMs rely on hard binary masking and discrete token assignments, which hinder the revision of early decisions and underutilize intermediate probabilistic representations. In this paper, we propose EvoToken-DLM, a novel diffusion-based language modeling approach that replaces hard binary masks with evolving soft token distributions. EvoToken-DLM enables a progressive transition from masked states to discrete outputs, supporting revisable decoding. To effectively support this evolution, we introduce continuous trajectory supervision, which aligns training objectives with iterative probabilistic updates. Extensive experiments across multiple benchmarks show that EvoToken-DLM consistently achieves superior performance, outperforming strong diffusion-based and masked DLM baselines. Project webpage: https://aim-uofa.github.io/EvoTokenDLM. |
| title | Beyond Hard Masks: Progressive Token Evolution for Diffusion Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2601.07351 |