XC-Cache: Cross-Attending to Cached Context for Efficient LLM Inference Paper • 2404.15420 • Published Apr 23 • 7