LLMAdBench: A Human Preference Benchmark for Advertising in LLM Responses.

arXiv:2609.32533v1 Announce Type: new Abstract: Inserting advertisements (ads) into consumer-facing LLM output is emerging as a new business model, but there is little shared evidence on how such ad insertion should be evaluated or how it affects user preferen