8 papers
Neurosymbolic LoRA: Why and When to Tune Weights vs. Rewrite Prompts
Kevin Wang, Neel P. Bhatt, Cong Liu +7
Large language models (LLMs) can be adapted either through numerical updates that alter model parameters or symbolic manipulations that work on discrete prompts or logical constrai…
LoX: Low-Rank Extrapolation Robustifies LLM Safety Against Fine-tuning
Gabriel J. Perin, Runjin Chen, Xuxi Chen +3
Large Language Models (LLMs) have become indispensable in real-world applications. However, their widespread adoption raises significant safety concerns, particularly in responding…
Learning Along the Arrow of Time: Hyperbolic Geometry for Backward-Compatible Representation Learning
Ngoc Bui, Menglin Yang, Runjin Chen +5
Backward compatible representation learning enables updated models to integrate seamlessly with existing ones, avoiding to reprocess stored data. Despite recent advances, existing…
SEAL: Steerable Reasoning Calibration of Large Language Models for Free
Runjin Chen, Zhenyu Zhang, Junyuan Hong +2
Large Language Models (LLMs), such as OpenAI's o1-series have demonstrated compelling capabilities for complex reasoning tasks via the extended chain-of-thought (CoT) reasoning mec…
More is Less: The Pitfalls of Multi-Model Synthetic Preference Data in DPO Safety Alignment
Yifan Wang, Runjin Chen, Bolian Li +7
Aligning large language models (LLMs) with human values is an increasingly critical step in post-training. Direct Preference Optimization (DPO) has emerged as a simple, yet effecti…
Extracting and Understanding the Superficial Knowledge in Alignment
Runjin Chen, Gabriel Jacob Perin, Xuxi Chen +5
Alignment of large language models (LLMs) with human values and preferences, often achieved through fine-tuning based on human feedback, is essential for ensuring safe and responsi…