Persimmon | humans&
persimmon.humansand.ai · 6,084 words · saved by 7 readers
A model of human behavior, and new ways to measure how faithfully AI can simulate people.
The first step in helping someone is understanding them; this is true for both humans & AI. People are unpredictable, behave differently in different social contexts, and change over time. Any model of human behavior needs to represent that variation. Today, we are releasing Persimmon, a large-scale model designed to realistically simulate how people behave in multi-turn, multi-user chat settings. This kind of model is often referred to as a "user model." Our user model is built to capture the distribution of human behavior within one person’s responses as well as across people, situations,…
saved by
related reading
- Persimmon v0.1 Model Cardpersimmon.humansand.ai
- The Persona Selection Model: Why AI Assistants might Behave like Humansalignment.anthropic.com
- Interaction Models: A Scalable Approach to Human-AI Collaboration - Thinking Machines Labthinkingmachines.ai
- User awareness in frontier modelstransluce.org
- [2505.23058] Be.FM: Open Foundation Models for Human Behaviorarxiv.org
- Announcing Transluce's Mental Health Evaluationtransluce.org
- Simulating human behavior | Similesimile.com
- [2505.10831] Creating General User Models from Computer Usearxiv.org
- [2606.06614] Re-Centering Humans in LLM Personalizationarxiv.org
- [2304.03442] Generative Agents: Interactive Simulacra of Human Behaviorarxiv.org
- trees are harlequins, words are harlequins - the voidnostalgebraist.tumblr.com
- [2608.19549] Generating Diverse Personas for User Simulators to Test Interview Dialogue Systemsarxiv.org