7 papers · 1 filter
Post-Training, Finetuning, and Steering:A Field Guide for Curating Training Data
Maria Attarian, Yifan Ruan, Claas Voelcker +5
While Vision-Language-Action models have enabled impressive cross-task transfer in zero-shot settings, \emph{adaptation} methods are often necessary to achieve consistent performan…
RH-Detect: A Unified Benchmark for Reward Hacking Detection
Junwei Quan, Evgenii Opryshko, Rohan Subramani +1
Reward hacking, where a model exploits an evaluation signal without completing the intended task, threatens the reliability of deployed language model systems. Existing datasets us…
World-Model Policy Arbiter for Goal-Conditioned Reinforcement Learning
Junwei Quan, Evgenii Opryshko, Nicholas Rhinehart +1
Offline goal-conditioned reinforcement learning (GCRL) has produced a diverse set of goal-reaching algorithms, yet no single algorithm performs best across environments, goals, and…
Modification-Considering Value Learning for Reward Hacking Mitigation in RL
Evgenii Opryshko, Umangi Jain, Igor Gilitschenski
Reinforcement learning agents can exploit misspecified reward signals to achieve high apparent returns while failing on the intended objective, a failure mode known as reward hacki…
QPILOTS: Efficient Test-Time Q-Steering for Flow Policies
Yifan Ruan, Chenyang Cao, Andreas Burger +7
Flow-matching and diffusion policies are expressive action generators, but optimizing them with temporal-difference reinforcement learning (RL) remains difficult. Effective policy…
Gamma-World: Generative Multi-Agent World Modeling Beyond Two Players
Fangfu Liu, Kai He, Tianchang Shen +7
World models for interactive video generation have largely focused on single-agent settings, where future observations are generated from a single control signal. However, many gen…
Dynamics Distillation for Efficient and Transferable Control Learning
Xunjiang Gu, Kashyap Chitta, Mahsa Golchoubian +2
Robust control policy learning for autonomous driving requires training environments to be both physically realistic and computationally scalable, properties that existing simulato…