Think Before Writing: Feature-Level Multi-Objective Optimization for Generative Citation Visibility

TL;DR

This paper introduces FeatGEO, a feature-level multi-objective optimization framework that enhances citation visibility in generative answer engines by optimizing interpretable webpage properties rather than direct text rewriting.

Contribution

FeatGEO offers a novel interpretable, high-level optimization approach that improves citation visibility while maintaining content quality, outperforming traditional token-level methods.

Findings

FeatGEO significantly improves citation visibility across multiple engines.

Document-level content properties influence citation behavior more than lexical edits.

Learned feature configurations generalize across different language model scales.

Abstract

Generative answer engines expose content through selective citation rather than ranked retrieval, fundamentally altering how visibility is determined. This shift calls for new optimization methods beyond traditional search engine optimization. Existing generative engine optimization (GEO) approaches primarily rely on token-level text rewriting, offering limited interpretability and weak control over the trade-off between citation visibility and content quality. We propose FeatGEO, a feature-level, multi-objective optimization framework that abstracts webpages into interpretable structural, content, and linguistic properties. Instead of directly editing text, FeatGEO optimizes over this feature space and uses a language model to realize feature configurations into natural language, decoupling high-level optimization from surface-level generation. Experiments on GEO-Bench across three generative engines demonstrate that FeatGEO consistently improves citation visibility while maintaining or improving content quality, substantially outperforming token-level baselines. Further analyses show that citation behavior is more strongly influenced by document-level content properties than by isolated lexical edits, and that the learned feature configurations generalize across language models of different scales.