TSPORec: Token Selection via Preference Optimization for LLM-Based Sequential Recommendation

2026-08-10

AI Summary

This work addresses the high computational overhead of existing large language model (LLM)-based sequential recommendation methods, which stems from processing full-length textual inputs, while naive truncation often discards critical information. To resolve this trade-off, we propose TSPORec, the first approach to integrate a preference optimization-driven token selection mechanism into LLM-based sequential recommendation. TSPORec employs a three-stage pipeline to precisely identify informative tokens and introduces a novel proxy reward mechanism to guide the selection process. This enables substantial reductions in computational cost without sacrificing—and indeed enhancing—recommendation accuracy. Extensive experiments on two benchmark datasets demonstrate that TSPORec consistently outperforms six strong baselines, achieving up to a 31.25% improvement in recommendation performance and a 63.4% gain in computational efficiency.

Abstract

Large Language Models (LLMs) have emerged as powerful tools for improving recommendation systems. The effectiveness of LLMs arises from their ability to harness rich textual information and their capacity to model heterogeneous user preferences based on users' interaction history. However, due to the large-scale and deep architectures, LLM-based sequential recommendation approaches generally incur high inference costs, resulting in a low return on investment. To mitigate this cost, many existing approaches resort to using only the first few tokens of item descriptions, which inadvertently discards valuable information contained in the full text, thereby leading to suboptimal recommendation performance. To address this limitation, we propose a novel Token Selection approach for Preference Optimization in LLM-based sequential Recommendation, i.e., TSPORec, which accurately pinpoints informative tokens throughout the entire textual content to improve recommendation performance. Specifically, we design a three-stage pipeline to select informative tokens and introduce a novel proxy reward to facilitate the implementation. TSPORec not only enhances recommendation performance but also improves computational efficiency. Extensive experiments across two models and datasets demonstrate the superb performance (up to 31.25%) and efficiency (up to 63.4%) of our approach compared with six baseline approaches. Code is available at https://github.com/WNQzhu/TSPORec.git.

Problem

• LLM-based sequential recommendation
• token selection
• inference cost
• recommendation performance
• textual information utilization

Innovation

• Token Selection
• Preference Optimization
• LLM-based Recommendation
• Sequential Recommendation
• Proxy Reward