patrickamadeus/Qwen2.5-VL-3B-Instruct-distill_lang Image-Text-to-Text • 4B • Updated about 5 hours ago • 1
Full Attention Strikes Back: Transferring Full Attention into Sparse within Hundred Training Steps Paper • 2605.16928 • Published 8 days ago • 82
SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training Paper • 2605.18719 • Published 6 days ago • 5
CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligence Paper • 2605.12882 • Published 11 days ago • 262
Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformers Paper • 2605.06169 • Published 17 days ago • 212
Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability Paper • 2604.06628 • Published Apr 8 • 325
Adam's Law: Textual Frequency Law on Large Language Models Paper • 2604.02176 • Published Apr 2 • 503