The End of Manual Decoding: Towards Truly End-to-End Language Models Paper • 2510.26697 • Published Oct 30 • 114
From Denoising to Refining: A Corrective Framework for Vision-Language Diffusion Model Paper • 2510.19871 • Published Oct 22 • 29
DeepWideSearch: Benchmarking Depth and Width in Agentic Information Seeking Paper • 2510.20168 • Published Oct 23 • 27
HSCodeComp: A Realistic and Expert-level Benchmark for Deep Search Agents in Hierarchical Rule Application Paper • 2510.19631 • Published Oct 22 • 27
ProLong Collection ProLong is a family of long-context models that are continued trained and supervised fine-tuned from Llama-3-8B, with a maximum context window of 512K • 7 items • Updated Mar 8 • 7