Training Dynamics of Multi-Head Softmax Attention for In-Context Learning: Emergence, Convergence, and Optimality (extended abstract).
Siyu Chen, Heejune Sheen, Tianhao Wang, Zhuoran Yang
Browse the full COLT paper archive.
Siyu Chen, Heejune Sheen, Tianhao Wang, Zhuoran Yang
Browse the full COLT paper archive.