When Simulation Lies: A Sim-to-Real Benchmark and Domain-Randomized RL Recipe for Tool-Use Agents

arXiv:2605.11928 · cs.AI · Submitted 2026-05-12 · Read on arXiv

cs.AI

Submitted: 2026-05-12

Updated: 2026-10-09

Comments: Accepted at NeurIPS 2026 (Evaluations and Datasets Track). Added acknowledgments. Dataset, code, and benchmark leaderboard are available at https://github.com/WillChow66/robustbench-tc-release.git and https://huggingface.co/spaces/willchow66/robustbench-tc-leaderboard

Code: https://github.com/WillChow66/robustbench-tc-release

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Accuracy — 4× in this paper · explained in 1 episode
• Benchmark — 4× in this paper · explained in 2 episodes
• Action Space — 1× in this paper · explained in 1 episode
• Partially observable Markov decision process (POMDP) — 1× in this paper · explained in 2 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Reinforcement Learning (RL) — 1× in this paper · explained in 25 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Sim-to-Real Gap — 1× in this paper · explained in 6 episodes

Sources

• Qwen2.5 Technical Report
• Qwen3 Technical Report
• TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use
• LoopTool: Closing the Data-Training Loop for Robust LLM Tool Calls
• MUA-RL: Multi-turn User-interacting Agent Reinforcement Learning for agentic tool use

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems