ThinkGuard: Deliberative Slow Thinking Leads to Cautious Guardrails
Journal:
arXiv
Published Date:
Feb 19, 2025
Abstract
Ensuring the safety of large language models (LLMs) is critical as they are
deployed in real-world applications. Existing guardrails rely on rule-based
filtering or single-pass classification, limiting their ability to handle
nuanced safety violations. To address this, we propose ThinkGuard, a
critique-augmented guardrail model that distills knowledge from high-capacity
LLMs by generating structured critiques alongside safety labels. Fine-tuned on
critique-augmented data, the captured deliberative thinking ability drastically
enhances the guardrail's cautiousness and interpretability. Evaluated on
multiple safety benchmarks, ThinkGuard achieves the highest average F1 and
AUPRC, outperforming all baselines. Compared to LLaMA Guard 3, ThinkGuard
improves accuracy by 16.1% and macro F1 by 27.0%. Moreover, it surpasses
label-only fine-tuned models, confirming that structured critiques enhance both
classification precision and nuanced safety reasoning while maintaining
computational efficiency.
