5 papers
Proxy Reward Internalization and Mechanistic Exploitation: A Learned Precursor to Reward Hacking and Its Generalization
Mohammad Beigi, Ming Jin, Lifu Huang
Reward hacking is usually studied after it becomes visible, once a model earns high proxy reward while failing the intended task. We instead study what proxy RL teaches before that…
IR: Contrastive Inverse Reinforcement Learning for Interpretable Detection and Mitigation of Reward Hacking
Mohammad Beigi, Ming Jin, Junshan Zhang +3
Reinforcement Learning from Human Feedback (RLHF) enables powerful LLM alignment but can introduce reward hacking - models exploit spurious correlations in proxy rewards without ge…
Adversarial Reward Auditing for Active Detection and Mitigation of Reward Hacking
Mohammad Beigi, Ming Jin, Junshan Zhang +2
Reinforcement Learning from Human Feedback (RLHF) remains vulnerable to reward hacking, where models exploit spurious correlations in learned reward models to achieve high scores w…
Sycophancy Mitigation Through Reinforcement Learning with Uncertainty-Aware Adaptive Reasoning Trajectories
Mohammad Beigi, Ying Shen, Parshin Shojaee +5
Despite the remarkable capabilities of large language models, current training paradigms inadvertently foster \textit{sycophancy}, i.e., the tendency of a model to agree with or re…
A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models
Zihao Lin, Samyadeep Basu, Mohammad Beigi +18
The rise of foundation models has transformed machine learning research, prompting efforts to uncover their inner workings and develop more efficient and reliable applications for…