TY  - RPRT
TI  - HarmRLVR: Weaponizing Verifiable Rewards for Harmful LLM Alignment
AU  - Yuexiao Liu
AU  - Lijun Li
AU  - Xingjun Wang
AU  - Jing Shao
PY  - 2026
UR  - https://arxiv.org/abs/2510.15499
ID  - 2510.15499
ER  -