@article{dsteerpreferencealignmenttechniqueslearn,
  title = {D-STEER - Preference Alignment Techniques Learn to Behave, not to Believe -- Beneath the Surface, DPO as Steering Vector Perturbation in Activation Space},
  author = {Samarth Raina and Saksham Aggarwal and Aman Chadha and Vinija Jain and Amitava Das},
  year = {2025},
  eprint = {2512.11838},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2512.11838},
}