Spectral Guardrails for Agents in the Wild: Detecting Tool Use Hallucinations via Attention Topology

📅 2026-02-08
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
This work addresses the challenge of hallucination in autonomous agents deployed in open environments when using external tools, where existing approaches lack reliable, training-free detection mechanisms. The authors propose a training-free spectral guardrail that identifies hallucinatory behavior by analyzing attention topology—specifically smoothness and entropy. Their findings reveal that spectral features from a single transformer layer can detect hallucinations with near-perfect accuracy, suggesting that hallucination fundamentally arises from abrupt shifts in the thermodynamic state of the model’s attention. The study also uncovers a “loud lies” phenomenon, wherein hallucinated outputs exhibit distinct spectral signatures. Evaluated on Llama 3.1 8B, the method achieves a 97.7% recall (98.2% using Layer 26 smoothness) and attains the best discriminative performance on Mistral 7B with an AUC of 0.900, demonstrating effective, label-free, and cross-model hallucination detection.

Technology Category

Application Category

📝 Abstract
Deploying autonomous agents in the wild requires reliable safeguards against tool use failures. We propose a training free guardrail based on spectral analysis of attention topology that complements supervised approaches. On Llama 3.1 8B, our method achieves 97.7\% recall with multi-feature detection and 86.1\% recall with 81.0\% precision for balanced deployment, without requiring any labeled training data. Most remarkably, we discover that single layer spectral features act as near-perfect hallucination detectors: Llama L26 Smoothness achieves 98.2\% recall (213/217 hallucinations caught) with a single threshold, and Mistral L3 Entropy achieves 94.7\% recall. This suggests hallucination is not merely a wrong token but a thermodynamic state change: the model's attention becomes noise when it errs. Through controlled cross-model evaluation on matched domains ($N=1000$, $T=0.3$, same General domain, hallucination rates 20--22\%), we reveal the ``Loud Liar''phenomenon: Llama 3.1 8B's failures are spectrally catastrophic and dramatically easier to detect, while Mistral 7B achieves the best discrimination (AUC 0.900). These findings establish spectral analysis as a principled, efficient framework for agent safety.
Problem

Research questions and friction points this paper is trying to address.

tool use hallucinations
autonomous agents
attention topology
spectral analysis
agent safety
Innovation

Methods, ideas, or system contributions that make the work stand out.

spectral analysis
attention topology
tool use hallucination
training-free detection
agent safety
🔎 Similar Papers