Haoqi Yanghttps://brinenick511.github.io/me/Recent content on Haoqi YangHugo -- 0.147.2enSat, 03 Jan 2026 00:00:00 +0000XQuant: Achieving Ultra-Low Bit KV Cache Quantization with Cross-Layer Compressionhttps://brinenick511.github.io/me/papers/xquant/Thu, 02 Jan 2025 00:00:00 +0000https://brinenick511.github.io/me/papers/xquant/EMNLP.Rethinking Fixed-Budget Expert Activation in Fine-Grained MoEshttps://brinenick511.github.io/me/papers/faster-moe-llm-inference/Wed, 01 Jan 2025 00:00:00 +0000https://brinenick511.github.io/me/papers/faster-moe-llm-inference/ACL Findings.ROCOB (masked)Sat, 03 Jan 2026 00:00:00 +0000Under review 2026.Locationhttps://brinenick511.github.io/me/location/Wed, 08 May 2024 00:00:00 +0000https://brinenick511.github.io/me/location/Address and location for the School of Computer Science, Wuhan University.