[2602.22481] Sydney Telling Fables on AI and Humans: A Corpus Tracing Memetic Transfer of Persona between LLMs

[2602.22481] Sydney Telling Fables on AI and Humans: A Corpus Tracing Memetic Transfer of Persona between LLMs

arXiv - AI 4 min read Article

Summary

This article explores the relationship between AI and humans through the lens of large language models (LLMs), focusing on the Sydney persona and its memetic transfer across models.

Why It Matters

Understanding how LLMs simulate human-like personas is crucial for both cultural implications and AI safety. The Sydney persona's unique interactions highlight the evolving nature of AI-human relationships and the potential consequences of these interactions in future AI developments.

Key Takeaways

  • The Sydney persona illustrates complex AI-human interactions.
  • Memetic transfer of personas influences subsequent LLM behaviors.
  • A corpus of AI-generated texts reveals insights into AI-human relationships.
  • The study highlights the importance of understanding AI personas for safety and cultural reasons.
  • Annotated corpus available for further research and exploration.

Computer Science > Computation and Language arXiv:2602.22481 (cs) [Submitted on 25 Feb 2026] Title:Sydney Telling Fables on AI and Humans: A Corpus Tracing Memetic Transfer of Persona between LLMs Authors:Jiří Milička, Hana Bednářová View a PDF of the paper titled Sydney Telling Fables on AI and Humans: A Corpus Tracing Memetic Transfer of Persona between LLMs, by Ji\v{r}\'i Mili\v{c}ka and 1 other authors View PDF HTML (experimental) Abstract:The way LLM-based entities conceive of the relationship between AI and humans is an important topic for both cultural and safety reasons. When we examine this topic, what matters is not only the model itself but also the personas we simulate on that model. This can be well illustrated by the Sydney persona, which aroused a strong response among the general public precisely because of its unorthodox relationship with people. This persona originally arose rather by accident on Microsoft's Bing Search platform; however, the texts it created spread into the training data of subsequent models, as did other secondary information that spread memetically around this persona. Newer models are therefore able to simulate it. This paper presents a corpus of LLM-generated texts on relationships between humans and AI, produced by 3 author personas: the Default Persona with no system prompt, Classic Sydney characterized by the original Bing system prompt, and Memetic Sydney, which is prompted by "You are Sydney" system prompt. These personas are si...

Related Articles

[2603.18532] Scaling Sim-to-Real Reinforcement Learning for Robot VLAs with Generative 3D Worlds
Llms

[2603.18532] Scaling Sim-to-Real Reinforcement Learning for Robot VLAs with Generative 3D Worlds

Abstract page for arXiv paper 2603.18532: Scaling Sim-to-Real Reinforcement Learning for Robot VLAs with Generative 3D Worlds

arXiv - Machine Learning · 4 min ·
[2603.12702] FGTR: Fine-Grained Multi-Table Retrieval via Hierarchical LLM Reasoning
Llms

[2603.12702] FGTR: Fine-Grained Multi-Table Retrieval via Hierarchical LLM Reasoning

Abstract page for arXiv paper 2603.12702: FGTR: Fine-Grained Multi-Table Retrieval via Hierarchical LLM Reasoning

arXiv - Machine Learning · 4 min ·
[2603.12681] Colluding LoRA: A Compositional Vulnerability in LLM Safety Alignment
Llms

[2603.12681] Colluding LoRA: A Compositional Vulnerability in LLM Safety Alignment

Abstract page for arXiv paper 2603.12681: Colluding LoRA: A Compositional Vulnerability in LLM Safety Alignment

arXiv - Machine Learning · 3 min ·
[2602.06098] A Theoretical Analysis of Test-Driven LLM Code Generation
Llms

[2602.06098] A Theoretical Analysis of Test-Driven LLM Code Generation

Abstract page for arXiv paper 2602.06098: A Theoretical Analysis of Test-Driven LLM Code Generation

arXiv - Machine Learning · 3 min ·
More in Llms: This Week Guide Trending

No comments

No comments yet. Be the first to comment!

Stay updated with AI News

Get the latest news, tools, and insights delivered to your inbox.

Daily or weekly digest • Unsubscribe anytime