Lijun Zhang

7 papers Nanjing University (NJU)

• Analyzing the Power of Chain of Thought through Memorization Capabilities
• Continuous Subspace Optimization for Continual Learning
• CoreGuard: Safeguarding Foundational Capabilities of LLMs Against Model Stealing in Edge Deployment
• Risk-aware Direct Preference Optimization under Nested Risk Measure
• SPACE: Noise Contrastive Estimation Stabilizes Self-Play Fine-Tuning for Large Language Models
• SeCon-RAG: A Two-Stage Semantic Filtering and Conflict-Free Framework for Trustworthy RAG
• Triplets Better Than Pairs: Towards Stable and Effective Self-Play Fine-Tuning for LLMs