@inproceedings{ijcai2026p83, title = {Detoxifying Large Language Models via Localized Feature Editing with Sparse Autoencoders}, author = {Shang, Yuhu and Cheng, Xiang and Wang, Dianyun and Ren, Yimeng and Luo, Xuexiong and Yang, Hang and Wu, Huijia and He, Zhaofeng}, booktitle = {Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, {IJCAI-26}}, publisher = {International Joint Conferences on Artificial Intelligence Organization}, editor = {Diego Calvanese}, pages = {735--743}, year = {2026}, month = {8}, note = {Main Track}, doi = {10.24963/ijcai.2026/83}, url = {https://doi.org/10.24963/ijcai.2026/83}, }