The paper introduces the Model Capacity Frontier (MCF), a framework that maps the empirical performance achievable by a given architecture across its possible configurations, particularly for large language model (LLM)-supported agent-based models (ABMs). Using a crowd simulation of university commencement events with about 12,000 agents, the authors demonstrate how MCF can identify which performance dimensions improve or remain limited as LLM layers are added and refined. The study shows that initial LLM integration enhances flow direction but worsens temporal occupancy and spatial distribution, while subsequent architectural tweaks recover these metrics and preserve flow gains.
arXiv:2602.00685v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used as simulated participants in social science experiments, but their behavior is often unstable an...
arXiv:2609.38816v1 Announce Type: new Abstract: While multi-agent and model collaboration algorithms gain traction to combine the strengths of diverse Large Language Models (LLMs), existing systems r...
arXiv:2507. 09788v3 Announce Type: replace-cross Abstract: Recent advances in Large Language Models (LLM) have led to a new class of autonomous agents, renewing and expanding interest in the area.
arXiv:2607. 24341v1 Announce Type: new Abstract: Recent studies use Large language models (LLMs) to simulate human opinions and decisions by prompting models with demographic, attitudinal, or persona-based descriptions.
The paper introduces a test‑time framework that blends Model Predictive Path Integral (MPPI) control with a dual‑branch experience memory for large language model agents. At each decision step, the agent generates multiple candidate actions, uses a world model to simulate their outcomes, and evaluates these predictions for task progress and risk to select the best action. The dual‑branch memory allows the planner to refine action generation and the world model to improve future observation predictions, leading to strong performance on ALFWorld and ScienceWorld benchmarks.
arXiv:2607. 25140v1 Announce Type: new Abstract: This paper studies the behavior of language models in a multi-agent crowd simulation, focusing on how affect propagates among agents that perceive and appraise one another.
arXiv:2507.19364v3 Announce Type: replace Abstract: The integration of Large Language Models (LLMs) into social simulation has generated considerable enthusiasm, but also raises substantial methodolo...
The paper introduces StressSim, an intent‑preserving adversarial user simulator designed to stress‑test multi‑turn agents by generating challenging yet valid user interactions. StressSim optimises high‑level adversarial plans for each task instance using a genetic prompt‑optimisation loop that leverages rollout traces, task‑completion signals, and validity feedback to reduce task success while keeping the user’s goal unchanged. Experiments on MultiWOZ, retail, and airline domains show that StressSim reveals significant robustness gaps in leading LLM agents—reducing tau‑bench pass^4 scores by up to 69%—while maintaining realistic conversation quality and goal preservation with a modest optimisation budget.
arXiv:2610.15788v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly deployed as teams, yet they fail in characteristically social ways: information holders stay silent,...
arXiv:2607. 17437v1 Announce Type: new Abstract: Large language model (LLM) agents offer a generative approach to simulating human behavior under conditions that may have few or no direct historical analogues, a common challenge in disaster and infrastructure-disruption planning.
Agentic World Modeling (AWM) redefines next-observation prediction as a multi-turn evidence-acquisition task, allowing agents to query a real environment safely and construct observations without mutating shared state. Integrated with trajectory-level test-time scaling, AWM enables parallel exploration over a single shared environment, outperforming single-shot simulation by 40.07% and trained world-model baselines by 12.36% on benchmarks such as τ-bench, AppWorld, and OfficeBench. This approach provides a practical foundation for safe and scalable agentic test-time scaling in shared, non-resettable environments.
While multi-agent and model collaboration algorithms gain traction to combine the strengths of diverse Large Language Models (LLMs), existing systems remain bottlenecked on pre-defined and hand-crafte...