Fetching the paper…

Faster Language Models with Better Multi-Token Prediction Using Tensor Decomposition · Around