{"id":"2608.28911","title":"SemKV: Semantic Mixed-Precision KV Cache Quantization Guided by the Quality Cliff for Long-Context LLM Inference","url":"https://codeofpaper.com/paper/2608.28911","arxiv_url":"https://arxiv.org/abs/2608.28911","published":"2026-08-28","code":"none","best":null,"most_starred_community":null,"implementations":[],"possible_matches":0}