5 papers
Quality-constrained Entropy Maximization Policy Optimization for LLM Diversity
Haihui Pan, Yuzhong Hong, Kaichen Zhang +4
In many large language model (LLM) alignment applications, users expect not only high-quality outputs but also substantial diversity. However, existing methods often face a fundame…
GVPO: Group Variance Policy Optimization for Large Language Model Post-Training
Kaichen Zhang, Yuzhong Hong, Junwei Bao +4
Post-training plays a crucial role in refining and aligning large language models to meet specific tasks and human preferences. While recent advancements in post-training technique…
RSPO: Risk-Seeking Policy Optimization for Pass@k and Max@k Metrics in Large Language Models
Kaichen Zhang, Shenghao Gao, Yuzhong Hong +6
Current large language model post-training optimizes a risk-neutral objective that maximizes expected reward, yet evaluation relies heavily on risk-seeking metrics like Pass@k (at…
Energy-Based Preference Model Offers Better Offline Alignment than the Bradley-Terry Preference Model
Yuzhong Hong, Hanshan Zhang, Junwei Bao +2
Since the debut of DPO, it has been shown that aligning a target LLM with human preferences via the KL-constrained RLHF loss is mathematically equivalent to a special kind of rewar…
Preference-Oriented Supervised Fine-Tuning: Favoring Target Model Over Aligned Large Language Models
Yuchen Fan, Yuzhong Hong, Qiushi Wang +3
Alignment, endowing a pre-trained Large language model (LLM) with the ability to follow instructions, is crucial for its real-world applications. Conventional supervised fine-tunin…