AI.info

A. Bochkov

Explore A. Bochkov on AI.info.

• Do Language Models Need a Trainable Input Embedding Table? Fixed Minimal Token Codes at 1.7B-Class Scale