← Search

Dueun Kim

2 accepted papers

2026

Rainbow Padding: Mitigating Early Termination in Instruction-Tuned Diffusion LLMs

ICLR 2026poster

Diffusion large language models (dLLMs) have emerged as a promising alternative to autoregressive models, offering flexible generation orders and strong performance on complex reasoning tasks. However, instruction-tuned dLLMs exhibit a critical vulnerability we term \<eos\> overflow: as allocated s…

Cited by 0SourceScholar
2025

Understanding and Mitigating Memorization in Generative Models via Sharpness of Probability Landscapes

ICML 2025spotlight

In this paper, we introduce a geometric framework to analyze memorization in diffusion models through the sharpness of the log probability density. We mathematically justify a previously proposed score-difference-based memorization metric by demonstrating its effectiveness in quantifying sharpness.…

Cited by 0SourcePDFScholar