LLM-Auction: Generative Auction towards LLM-Native Advertising

Source: arXiv Authors: Chujie Zhao, Qun Hu, Shiping Song, Dagui Chen, Han Zhu, Jian Xu, Bo Zheng Date: 2025-12-11 Primary category: cs.GT All categories: cs.GT, cs.AI, cs.LG

Abstract

LLM-Auction extends the mechanism-design line by coupling allocation and generation instead of treating them as separate stages. The application is LLM-native advertising, but the underlying point is broader: the paper learns an allocation policy that accounts for externalities directly in the generation loop and then studies the resulting incentive properties. That makes it relevant as an adjacent example of bid-aware coordination for systems where outputs, not just fixed slots, are being allocated.