The Cold-Start Safety Gap in LLM Agents

arXiv:2606.07867 · cs.CL · Submitted 2026-06-05 · Read on arXiv

cs.CL

Submitted: 2026-06-05

Updated: 2026-08-27

Code: https://github.com/Trustworthy-MLLab/Agent-Cold-Start-Safety-Gap

Terminology

• Benchmark — 65× in this paper · explained in 2 episodes
• Agentic Task — 42× in this paper · explained in 1 episode
• Agentic Tasks — 38× in this paper · explained in 1 episode
• Ablation — 18× in this paper · explained in 1 episode
• Fine-Tuning — 7× in this paper · explained in 5 episodes
• Accuracy — 6× in this paper · explained in 1 episode
• Alignment — 5× in this paper · explained in 1 episode
• Safety Over Depth for Agents — 4× in this paper
• Agent Safety Bench — 3× in this paper
• Computational overhead — 3× in this paper · explained in 2 episodes
• Prompt Engineering — 2× in this paper · explained in 4 episodes
• Adversarial Attacks — 1× in this paper · explained in 1 episode
• Adversarial Context — 1× in this paper · explained in 1 episode
• Agentic Evaluation — 1× in this paper · explained in 1 episode
• Benchmarking — 1× in this paper · explained in 3 episodes
• In-Context Learning — 1× in this paper · explained in 4 episodes
• In-Context Learning (ICL) — 1× in this paper · explained in 5 episodes
• LLM judge — 1× in this paper · explained in 2 episodes
• Representation Engineering — 1× in this paper · explained in 2 episodes

Sources

• LLM Defenses Are Not Robust to Multi-Turn Human Jailbreaks Yet
• LLM Agents Already Know When to Call Tools -- Even Without Reasoning
• The Llama 3 Herd of Models
• Representation Engineering: A Top-Down Approach to AI Transparency
• Qwen3.5-Omni Technical Report
• Voyager: An Open-Ended Embodied Agent with Large Language Models
• Qwen3 Technical Report
• Auto-GPT for Online Decision Making: Benchmarks and Additional Opinions
• ReAct: Synergizing Reasoning and Acting in Language Models
• AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models
• Agent-SafetyBench: Evaluating the Safety of LLM Agents

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering