TY  - RPRT
TI  - Agent-RewardBench: Towards a Unified Benchmark for Reward Modeling across Perception, Planning, and Safety in Real-World Multimodal Agents
AU  - Tianyi Men
AU  - Zhuoran Jin
AU  - Pengfei Cao
AU  - Yubo Chen
AU  - Kang Liu
AU  - Jun Zhao
PY  - 2025
UR  - https://arxiv.org/abs/2506.21252
ID  - 2506.21252
ER  -