pith. sign in

Guangxuan Xiao

Identifiers

  • name variant Guangxuan Xiao 0.60 · backfill

Papers (20)

  1. SparDA: Sparse Decoupled Attention for Efficient Long-Context LLM Inference cs.CL · 2026 · author #2
  2. BLASST: Dynamic BLocked Attention Sparsity via Softmax Thresholding cs.CL · 2025 · author #6
  3. Four Over Six: More Accurate NVFP4 Quantization with Adaptive Block Scaling cs.CL · 2025 · author #3
  4. StreamingVLM: Real-Time Understanding for Infinite Video Streams cs.CV · 2025 · author #2
  5. XAttention: Block Sparse Attention with Antidiagonal Scoring cs.CL · 2025 · author #2
  6. LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention cs.CL · 2025 · author #5
  7. DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads cs.CL · 2024 · author #1
  8. Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference cs.CL · 2024 · author #4
  9. QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving cs.CL · 2024 · author #5
  10. Retrieval Head Mechanistically Explains Long-Context Factuality cs.CL · 2024 · author #3
  11. BitDelta: Your Fine-Tune May Only Be Worth One Bit cs.LG · 2024 · author #2
  12. InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory cs.CL · 2024 · author #4
  13. Efficient Streaming Language Models with Attention Sinks cs.CL · 2023 · author #1
  14. AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration cs.CL · 2023 · author #7
  15. FastComposer: Tuning-Free Multi-Subject Image Generation with Localized Attention cs.CV · 2023 · author #1
  16. Sparse and Local Networks for Hypergraph Reasoning cs.LG · 2023 · author #1
  17. Offsite-Tuning: Transfer Learning without Full Model cs.CL · 2023 · author #1
  18. FreshGNN: Reducing Memory Access via Stable Historical Embeddings for Graph Neural Network Training cs.LG · 2023 · author #4
  19. SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models cs.CL · 2022 · author #1
  20. Red Alarm for Pre-trained Models: Universal Vulnerability to Neuron-Level Backdoor Attacks cs.CL · 2021 · author #2

Mentions

  • 2405.04532 #5 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2502.14866 #5 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2503.16428 #2 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2402.10193 #2 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2406.10774 #4 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2402.04617 #4 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2404.15574 #3 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2309.17453 #1 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2211.10438 #1 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2301.07482 #4 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2101.06969 #2 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2305.10431 #1 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2303.05496 #1 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2302.04870 #1 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2606.04511 #2 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2410.10819 #1 · arxiv_oai · confidence 0.70 Guangxuan Xiao
  • 2510.09608 #2 · arxiv_oai · confidence 0.70 Guangxuan Xiao

Frequent Coauthors