DADSA: Dual-Side Adaptive Deep Safety Alignment for Large Language Models.

Kunlin Li|Yabin Peng|Chenyu Zhou|Fan Zhang|Jiangtao Ma|Yaqiong Qiao|Wei Huang|Xinghao Wang


Anthology ID:DBLP:journals/ipm/LiPZZMQHW26
Year:2026
Venue:Information Processing and Management
Pages:104907
URL:https://doi.org/10.1016/J.IPM.2026.104907
DOI:https://doi.org/10.1016/J.IPM.2026.104907
DBLP:journals/ipm/LiPZZMQHW26
BibTeX:
@article{li-2026-dadsa, author = {Kunlin Li and Yabin Peng and Chenyu Zhou and Fan Zhang and Jiangtao Ma and Yaqiong Qiao and Wei Huang and Xinghao Wang}, title = {{DADSA: Dual-Side Adaptive Deep Safety Alignment for Large Language Models}}, journal = {Information Processing and Management}, volume = {63}, number = {8}, pages = {104907}, year = {2026}, url = {https://doi.org/10.1016/J.IPM.2026.104907}, doi = {https://doi.org/10.1016/J.IPM.2026.104907} }