ThreadBench: Benchmarking Realism in Agentic Simulation of Real-world Discussions

arXiv:2606.14715 · cs.MA, cs.AI, cs.SI · Submitted 2026-05-10 · Read on arXiv

cs.MA, cs.AI, cs.SI

Submitted: 2026-05-10

Updated: 2026-10-10

Code: https://github.com/unitaryai/detoxify

Terminology

• Benchmark — 2× in this paper · explained in 2 episodes
• Benchmarking — 1× in this paper · explained in 4 episodes

Sources

• Y Social: an LLM-powered Social Media Digital Twin
• SimBench: Benchmarking the Ability of Large Language Models to Simulate Human Behaviors
• BERTScore: Evaluating Text Generation with BERT

Related papers

• Highway Congestion Reduction through Reinforcement Learning Based Eulerian Headway Control
• ORACLE-SWE: Quantifying the Contribution of Critical Information Signals on SWE Agents
• You Only Align Once: Propagating Cooperative Behaviors in Multi-Agent Systems through Seed Agents
• Deny Without Disabling: Authorization-Paired Evaluation and Control for Multi-Agent Systems
• MA-SAPO: Multi-Agent Reasoning for Score-Aware Prompt Optimization
• PeroMAS: A Multi-agent System of Perovskite Material Discovery