@article{kvrunaheadscalablecausalllminferenceby, title = {KV-Runahead: Scalable Causal LLM Inference by Parallel Key-Value Cache Generation}, author = {Minsik Cho and Mohammad Rastegari and Devang Naik}, year = {2024}, eprint = {2405.05329}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2405.05329v2}, }