Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories

A paper submitted to arXiv on 26 May 2026 found that 11 words appear in 88.3% of LLM-generated stories, based on 20,000 stories sampled from four current models across five prompts. The recurring tokens include names (Elias, Mara, Elara), settings (lighthouses) and professions (clockmaker, librarian), and the paper reports little difference between models; these tokens are rare in published literature and pre-training data but appear in preference data likely used by all current models. The authors attribute the pattern to the disproportionate impact of small datasets combined with powerful alignment algorithms.

Computer Science Computation and Language Submitted on 26 May 2026 Title:Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories View PDF https://arxiv.org/pdf/2605.26492 HTML experimental https://arxiv.org/html/2605.26492v1 Abstract:LLM-generated stories are a popular use case, but they show very low variability. We sample 20,000 total stories from four current models using five prompts. We find that 11 words occur in 88.3% of generated stories, with little difference between models. These words include names Elias, Mara, Elara , settings lighthouses , and professions clockmaker, librarian . These tokens do not often occur in published literature nor pre-training data, but they are found in preference data that is likely to have been used by all current models. Surprisingly, these "lighthouse" stories are infrequent when compared with the average post-training story, much of which contains references to copyrighted characters or adult content. This result demonstrates the potentially disproportionate impact of small datasets combined with powerful alignment algorithms. Current browse context: cs.CL References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .