CLAug 5, 2025

CTR-Sink: Attention Sink for Language Models in Click-Through Rate Prediction

arXiv:2508.03668v12 citationsh-index: 9Has Code
Originality Incremental advance
AI Analysis

This addresses a structural gap in recommendation systems for better modeling user behavior sequences, though it is incremental as it builds on existing attention sink theory.

The paper tackles the problem of semantic fragmentation in using language models for click-through rate prediction by proposing CTR-Sink, a framework that introduces behavior-level attention sinks, resulting in improved prediction performance validated on industrial and open-source datasets.

Click-Through Rate (CTR) prediction, a core task in recommendation systems, estimates user click likelihood using historical behavioral data. Modeling user behavior sequences as text to leverage Language Models (LMs) for this task has gained traction, owing to LMs' strong semantic understanding and contextual modeling capabilities. However, a critical structural gap exists: user behavior sequences consist of discrete actions connected by semantically empty separators, differing fundamentally from the coherent natural language in LM pre-training. This mismatch causes semantic fragmentation, where LM attention scatters across irrelevant tokens instead of focusing on meaningful behavior boundaries and inter-behavior relationships, degrading prediction performance. To address this, we propose $\textit{CTR-Sink}$, a novel framework introducing behavior-level attention sinks tailored for recommendation scenarios. Inspired by attention sink theory, it constructs attention focus sinks and dynamically regulates attention aggregation via external information. Specifically, we insert sink tokens between consecutive behaviors, incorporating recommendation-specific signals such as temporal distance to serve as stable attention sinks. To enhance generality, we design a two-stage training strategy that explicitly guides LM attention toward sink tokens and a attention sink mechanism that amplifies inter-sink dependencies to better capture behavioral correlations. Experiments on one industrial dataset and two open-source datasets (MovieLens, Kuairec), alongside visualization results, validate the method's effectiveness across scenarios.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes