Prompt template choice creates an evaluation illusion in dLLMs where parallel decoding appears competitive, but experiments show these methods consistently underperform single-token decoding across varied settings.
Lookahead path likelihood optimization for diffusion llms.arXiv preprint arXiv:2602.03496, 2026
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Understanding Evaluation Illusion in Diffusion Large Language Models
Prompt template choice creates an evaluation illusion in dLLMs where parallel decoding appears competitive, but experiments show these methods consistently underperform single-token decoding across varied settings.