FeedSources
⌘K
HA

Hao AI Lab Publications

6 articles total

Go to source

This page showcases Hao AI Lab's extensive publication record at top-tier conferences like ICLR, NeurIPS, and ICML, highlighting breakthrough research in LLM serving, speculative decoding, and distributed machine learning parallelism.

Go to source
  •  d3LLM: Ultra-Fast Diffusion LLM using Pseudo-Trajectory Distillation
  •  Efficient Long-context Language Model Training by Core Attention Disaggregation
  •  Stronger Together: On-Policy Reinforcement Learning for Collaborative LLMs
  •  Deep Think with Confidence
  •  Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing
  •  LMGame-Bench: How Good are LLMs at Playing Games?