Parasitic Co-Denoising: Unlocking 3D Human Motion Generation in a Frozen Video Diffusion ModelResearchOct 5, 2026
Building Interpretable Feature Representations for Resume-Vacancy Matching by Distilling Production LLM SignalsResearchOct 5, 2026
Investigating the Role of Reasoning-Language Alignment in Monolingual Retrieval-Augmented GenerationResearchOct 5, 2026
Where to Look Is Not How to Fix: Pre-Denoising Diagnostics and Modality-Dependent Control in Diffusion CompositionResearchOct 5, 2026
NegT2IBench: When Negation Changes the Picture. A Polarity Benchmark for Text-to-Image ModelsResearchOct 5, 2026
LUMOS: Tracing Parametric Knowledge from Training Data to Behavioral Outputs in LLMsResearchOct 5, 2026