Multimodal Reward Hacking in Reinforcement Learning
英文标题：Multimodal Reward Hacking in Reinforcement Learning
arXiv:2607.09492v1 Announce Type: new Abstract: Reinforcement learning (RL) is increasingly used to align multimodal large language models (MLLMs), but higher rewards do not always imply better task performance. This risk is amplified when visual evidence is evaluated by text-only or weakly grounde
打开原文相关 2.4 · 机会 0
2.6 KB
2.1 KB
640 B
612 B
380 B
9.1 KB
32.7 KB
22.8 KB
24.2 KB
34.7 KB
中文
arXiv:2607.09492v1 Announce Type: new Abstract: Reinforcement learning (RL) is increasingly used to align multimodal large language models (MLLMs), but higher rewards do not always imply better task performance. This risk is amplified when visual evidence is evaluated by text-only or weakly grounde
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
BibTeX formatted citation
×