GHOST 2.0: generative high-fidelity one shot transfer of heads Paper • 2502.18417 • Published 13 days ago • 63
LLM-Microscope: Uncovering the Hidden Role of Punctuation in Context Memory of Transformers Paper • 2502.15007 • Published 18 days ago • 160
Cramming 1568 Tokens into a Single Vector and Back Again: Exploring the Limits of Embedding Space Capacity Paper • 2502.13063 • Published 20 days ago • 65