Review-LLM: Harnessing Large Language Models for Personalized Review Generation

TL;DR

<3-5 sentence high-level summary> The paper tackles personalized review generation in recommender systems by leveraging LLMs with a history-aware prompt that encodes user-item interactions and ratings. It introduces Review-LLM, which uses supervised fine-tuning and LoRA-based parameter-efficient training to tailor reviews for target items. Experiments on five Amazon datasets show that Review-LLM outperforms open- and closed-source baselines in automatic and human evaluations, particularly when rating information is included. The approach demonstrates that integrating user context into prompts can mitigate LLM politeness and improve sentiment-aligned, personalized explanations for recommendations.

Abstract

Product review generation is an important task in recommender systems, which could provide explanation and persuasiveness for the recommendation. Recently, Large Language Models (LLMs, e.g., ChatGPT) have shown superior text modeling and generating ability, which could be applied in review generation. However, directly applying the LLMs for generating reviews might be troubled by the ``polite'' phenomenon of the LLMs and could not generate personalized reviews (e.g., negative reviews). In this paper, we propose Review-LLM that customizes LLMs for personalized review generation. Firstly, we construct the prompt input by aggregating user historical behaviors, which include corresponding item titles and reviews. This enables the LLMs to capture user interest features and review writing style. Secondly, we incorporate ratings as indicators of satisfaction into the prompt, which could further improve the model's understanding of user preferences and the sentiment tendency control of generated reviews. Finally, we feed the prompt text into LLMs, and use Supervised Fine-Tuning (SFT) to make the model generate personalized reviews for the given user and target item. Experimental results on the real-world dataset show that our fine-tuned model could achieve better review generation performance than existing close-source LLMs.

Figures (3)

• Figure 1: An example of input prompt for Review-LLM.
• Figure 2: Human Evaluation. The bar is the mean of the model performance, and the error bar represents the max and min accuracy of the model.
• Figure 3: Case Study.