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

TL;DR

SEAT is a novel fine-tuning method that enhances knowledge adaptation in LLMs while preserving the model's ability to abstain from answering unknown queries, crucial for safety in high-stakes applications.

Contribution

SEAT introduces a sparse tuning and entity-perturbed KL regularization approach that maintains epistemic abstention without requiring alignment data or post-hoc adjustments.

Findings

SEAT improves abstention on unknown queries by 18%-101%.

SEAT retains near-perfect knowledge acquisition.

SEAT produces coherent, context-aware abstentions.

Abstract

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. SEAT combines sparse tuning, which constrains global activation drift, with entity-perturbed KL regularization, which sharpens local epistemic boundaries and prevents spillover to neighboring knowledge. Crucially, SEAT requires no alignment data, explicit boundary probing, or post-hoc re-alignment, making it attractive for lightweight and privacy-sensitive adaptation. Across models and datasets, SEAT improves human-evaluated abstention on unknown queries by 18%-101% over the strongest baseline while retaining near-perfect target knowledge acquisition, and produces coherent, context-aware abstentions after tuning. Further analyses show that both components are essential, that SEAT more cleanly separates known from unknown queries in representation space, and that it preserves downstream utility. These results identify preservation of epistemic abstention as a core objective for safe knowledge adaptation.