Large language models suffer from their own output: An analysis of the self-consuming training loop

M Briesch, D Sobania, F Rothlauf - arXiv preprint arXiv:2311.16822, 2023 - arxiv.org
Large language models (LLM) have become state of the art in many benchmarks and
conversational LLM applications like ChatGPT are now widely used by the public. Those …

[PDF][PDF] The Many Faces of AI Art: Self-Poisoning Generative Models

A Alazaroaie - 2024 - repository.tudelft.nl
As AI-generated content becomes more prevalent, the risk of generative models consuming
and regenerating their own outputs in a self-consuming loop increases. This study explores …