arXiv cs.CL· Haibo Jin, Huimin Zeng, Yifeng Luo, Haohan Wang·· 13 小时前AI 评分42
CORE：通过生成引擎优化控制 LLM 搜索的输出排名
Controlling Output Rankings in LLM-Based Search via Generative Engine Optimization
AI 导读
研究者提出 CORE 方法，通过向搜索引擎返回的内容追加字符串、推理、评论三类优化内容，操控 LLM 搜索的输出排名。在 GPT-4o、Gemini-2.5、Claude-4 和 Grok-3 四个模型上，CORE 在 15 个商品类目平均达到 91.4% @Top-5、86.6% @Top-3、80.3% @Top-1 的提升成功率。
正文 · 原文
Abstract:The way customers search for and choose products is changing with the rise of large language models (LLMs). LLM-based search, or generative engines, provides direct product recommendations to users, rather than traditional online search results that require users to explore options themselves. However, these recommendations are strongly influenced by the initial retrieval order of LLMs, which disadvantages small businesses and independent creators by limiting their visibility.
In this work, we propose CORE, an optimization method that \textbf{C}ontrols \textbf{O}utput \textbf{R}ankings in g\textbf{E}nerative Engines for LLM-based search. Since the LLM's interactions with the search engine are black-box, CORE targets the content returned by search engines as the primary means of influencing output rankings. Specifically, CORE optimizes retrieved content by appending strategically designed optimization content to steer the ranking of outputs. We introduce three types of optimization content: string-based, reasoning-based, and review-based, demonstrating their effectiveness in shaping output rankings. To evaluate CORE in realistic settings, we introduce ProductBench, a large-scale benchmark with 15 product categories and 200 products per category, where each product is associated with its top-10 recommendations collected from Amazon's search interface.
Extensive experiments on four LLMs with search capabilities (GPT-4o, Gemini-2.5, Claude-4, and Grok-3) demonstrate that CORE achieves an average Promotion Success Rate of \textbf{91.4\% @Top-5}, \textbf{86.6\% @Top-3}, and \textbf{80.3\% @Top-1}, across 15 product categories, outperforming existing ranking manipulation methods while preserving the fluency of optimized content.
| Comments: | 23 pages |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2602.03608 [cs.CL] |
| (or arXiv:2602.03608v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2602.03608 arXiv-issued DOI via DataCite |
Submission history
From: Haibo Jin [view email]
[v1]
Tue, 3 Feb 2026 14:59:48 UTC (476 KB)
[v2]
Wed, 21 Oct 2026 05:18:42 UTC (475 KB)
来源：arXiv cs.CL · arxiv.org