RetrievalAttention: Accelerating Long-Context LLM Inference via Vector Retrieval

Jun-17-2026, 04:41:09 GMT–Neural Information Processing Systems 

Transformer-based Large Language Models (LLMs) have become increasingly important.