Showing cs.AIShow all
2 papers · 1 filter
cs.AI2026
Beyond Routing Weights: Faithful Response-Level Interpretation of Mixture-of-Experts Reward Models via Contribution Contrast
Yifan Wang, Jinyi Mu, Mayank Jobanputra +4
Reward models are central to learning from human preferences, yet identifying what drives their predictions remains challenging. Recent sparse Mixture-of-Experts (MoE) reward model…
cs.AI2024
RSA-Control: A Pragmatics-Grounded Lightweight Controllable Text Generation Framework
Yifan Wang, Vera Demberg
Despite significant advancements in natural language generation, controlling language models to produce texts with desired attributes remains a formidable challenge. In this work,…