PulseAugur
EN
LIVE 10:33:28

New foundation models aim to simulate human behavior at scale

Researchers have introduced OdysSim, a new framework for developing foundation models designed to simulate human behavior. This initiative includes a large corpus of 21.4 million interactions and a benchmark called SOUL-Index, which unifies 23 tasks across five capability axes. The resulting 8B parameter model, OSim, demonstrates strong performance, ranking first on 8 tasks and showing human-like output quality, even transferring zero-shot to out-of-distribution user simulation. AI

IMPACT This research could advance the development of more realistic AI simulators for evaluation and social simulation, potentially improving human-AI interaction.

RANK_REASON The cluster describes a new research paper introducing a novel framework and model for simulating human behavior.

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New foundation models aim to simulate human behavior at scale

COVERAGE [2]

  1. arXiv cs.AI TIER_1 English(EN) · Xuhui Zhou, Weiwei Sun, Weihua Du, Jiarui Liu, Haojia Sun, Qianou Ma, Tongshuang Wu, Yiming Yang, Maarten Sap ·

    OdysSim: Building Foundation Models for Human Behavior Simulation

    arXiv:2606.14199v1 Announce Type: cross Abstract: Large language models are increasingly deployed as human simulators for interactive evaluation and social simulation. Yet helpfulness-driven post-training pulls them toward a homogeneous, overly agreeable assistant register, creat…

  2. arXiv cs.AI TIER_1 English(EN) · Maarten Sap ·

    OdysSim: Building Foundation Models for Human Behavior Simulation

    Large language models are increasingly deployed as human simulators for interactive evaluation and social simulation. Yet helpfulness-driven post-training pulls them toward a homogeneous, overly agreeable assistant register, creating a behavioral Sim2Real gap. We present OdysSim,…