AI.info

Xiaodong Gu

Explore Xiaodong Gu on AI.info.

• EarlyEval: Cheaper Agent Evaluation via Early Outcome Prediction
• SRPO: Self-Reflective Policy Optimization for Long-Horizon Reasoning
• SWE-Bench ProMax: Benchmarking Agents on Large-Scale Multilingual Code Refactoring
• SWE-Pruner Pro: The Coder LLM Already Knows What to Prune
• Seeing is Coding: On the Effectiveness of Vision Language Models in Code Understanding
• Reasoning in Trees: Improving Retrieval-Augmented Generation for Multi-Hop Question Answering
• GlimpRouter: Efficient Collaborative Inference by Glimpsing One Token of Thoughts
• Progressive Supernet Training for Efficient Visual Autoregressive Modeling
• Robust Preference Alignment via Directional Neighborhood Consensus