Y. Nam, J. Kim, and J. Jeong, “Learning from the Undesirable: Robust Adaptation of Language Models Without Forgetting”, AAAI, vol. 40, no. 38, pp. 32537–32545, Mar. 2026.