Residual Context Diffusion (RCD): Repurposing discarded signals as structured priors for high-performance reasoning in dLLMs.
Yuezhou Hu
yuezhouhu
AI & ML interests
My research interests include efficient machine learning, particularly efficient training and inference.
Recent Activity
updated a dataset 5 days ago
yuezhouhu/dolma3_dolmino_mix-100B-1125-tokenized-shuffled published a dataset 5 days ago
yuezhouhu/dolma3_dolmino_mix-100B-1125-tokenized-shuffled updated a dataset 7 days ago
yuezhouhu/dolma3_dolmino_mix-10B-1025-tokenized-shuffledOrganizations
models 12
yuezhouhu/RCD-LLaDA-8B-Instruct
1B • Updated • 17 • 1
yuezhouhu/SeqD-LLaDA-8B-Instruct
1B • Updated • 21
yuezhouhu/RCD-SDAR-8B-b64-Thinking
8B • Updated • 21
yuezhouhu/RCD-SDAR-8B-b32-Thinking
8B • Updated • 12
yuezhouhu/RCD-SDAR-4B-b64-Thinking
4B • Updated • 14
yuezhouhu/RCD-SDAR-4B-b32-Thinking
4B • Updated • 15
yuezhouhu/SeqD-SDAR-8B-b64-Thinking
8B • Updated • 10
yuezhouhu/SeqD-SDAR-8B-b32-Thinking
8B • Updated • 12
yuezhouhu/SeqD-SDAR-4B-b64-Thinking
4B • Updated • 11
yuezhouhu/SeqD-SDAR-4B-b32-Thinking
4B • Updated • 17