Reproduce “LLM-Auction: Generative Auction towards LLM-Native Advertising” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Chujie Zhao, Qun Hu, Shiping Song, Dagui Chen, Han Zhu, Jian Xu, Bo Zheng
LLM-Auction integrates ad allocation into answer generation, training the response model to favor complete responses that balance bid-weighted predicted clicks with penalties for excessive ads and formatting errors. In a simulation with fixed, single-round bids and shared first-price evaluation, it achieved 170.86 revenue per query versus MOSAIC’s 106.56, while inserting fewer ads; this is not evidence from users or markets.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.