Adaptive Activation Steering: A Tuning-Free LLM Truthfulness Improvement Method for Diverse Hallucinations Categories.

@inproceedings{wang-2025-adaptive, author = {Tianlong Wang and Xianfeng Jiao and Yinghao Zhu and Zhongzhi Chen and Yifan He and Xu Chu and Junyi Gao and Yasha Wang and Liantao Ma}, editor = {Guodong Long and Michale Blumestein and Yi Chang and Liane Lewin-Eytan and Zi Huang and Elad Yom-Tov}, title = {{Adaptive Activation Steering: A Tuning-Free LLM Truthfulness Improvement Method for Diverse Hallucinations Categories}}, booktitle = {{Proceedings of the ACM on Web Conference 2025, WWW 2025, Sydney, NSW, Australia, 28 April 2025- 2 May 2025}}, pages = {2562--2578}, publisher = {ACM}, year = {2025}, url = {https://doi.org/10.1145/3696410.3714640}, doi = {https://doi.org/10.1145/3696410.3714640} }