SEAT: Sparse Entity-Aware Tuning for Knowledge Adaptation while Preserving Epistemic Abstention

ArXi:2506.14387v3 Announce Type: replace Adapting LLMs with new knowledge is increasingly important, but standard fine-tuning often erodes aligned epistemic abstention: the ability to acknowledge when the model does not know. This failure mode is especially concerning in high-stakes settings, where abstention is a critical safeguard against hallucination. We present SEAT, a preventive fine-tuning method that preserves epistemic abstention while maintaining strong knowledge acquisition.