Guangxuan Xiao
Source-listed: Member of Technical Staff · Thinking Machines Lab
Record snapshot: Sep 3, 2026. Coverage and affiliations may be incomplete or historical.
Selected work
5- Efficient Streaming Language Models with Attention SinksSaved credit: First author · 2024Explore people connected to this work →
- SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language ModelsSaved credit: Contributor · 2023Explore people connected to this work →
- DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming HeadsSaved credit: First author · 2025Explore people connected to this work →
- XAttention: Block Sparse Attention with Antidiagonal ScoringSaved credit: First author · 2025Explore people connected to this work →
- StreamingVLM: Real-Time Understanding for Infinite Video StreamsSaved credit: First author · 2026Explore people connected to this work →