Cramming 1568 Tokens into a Single Vector and Back Again: Exploring the Limits of Embedding Space Capacity Paper • 2502.13063 • Published 23 days ago • 66
Soundwave: Less is More for Speech-Text Alignment in LLMs Paper • 2502.12900 • Published 23 days ago • 77
How Much Knowledge Can You Pack into a LoRA Adapter without Harming LLM? Paper • 2502.14502 • Published 21 days ago • 85
SuperGPQA: Scaling LLM Evaluation across 285 Graduate Disciplines Paper • 2502.14739 • Published 21 days ago • 97
Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention Paper • 2502.11089 • Published 25 days ago • 142