Yongbin Li

7

Papers

427

Total Citations

Papers (7)

Preference Ranking Optimization for Human Alignment

StructRAG: Boosting Knowledge Intensive Reasoning of LLMs via Inference-time Hybrid Information Structurization

On the Role of Attention Heads in Large Language Model Safety

Scaling Offline Model-Based RL via Jointly-Optimized World-Action Model Pretraining

Aligning Logits Generatively for Principled Black-Box Knowledge Distillation

Semantically-Shifted Incremental Adapter-Tuning is A Continual ViTransformer

Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch