AI RESEARCH
GeoLAN: Geometric Learning of Latent Explanatory Directions in Large Language Models
arXiv CS.LG
•
ArXi:2603.19460v1 Announce Type: new Large language models (LLMs) nstrate strong performance, but they often lack transparency. We