GuardReasoner: Towards Reasoning-based LLM Safeguards
yueliu1999
Abstract
As LLMs increasingly impact safety-critical applications, ensuring their safety using guardrails remains a key challenge. This paper proposes GuardReasoner, a new safeguard for LLMs, by guiding the guard model to learn to reason. Concretely, we first create the GuardReasonerTrain dataset, which consists of 127K samples with 460K detailed reasoning steps. Then, we introduce reasoning SFT to unlock the reasoning capability of guard models. In addition, we present hard sample DPO to further strengthen their reasoning ability. In this manner, GuardReasoner achieves better performance, explainability, and generalizability. Extensive experiments and analyses on 13 benchmarks of 3 guardrail tasks demonstrate its superiority. Remarkably, GuardReasoner 8B surpasses GPT-4o+CoT by 5.74% and LLaMA Guard 3 8B by 20.84% F1 score on average. We release the training data, code, and models with different scales (1B, 3B, 8B) of GuardReasoner : https://github.com/yueliu1999/GuardReasoner/.
Community
As LLMs increasingly impact safety-critical applications, ensuring their safety using guardrails remains a key challenge. This paper proposes GuardReasoner, a new safeguard for LLMs, by guiding the guard model to learn to reason. Concretely, we first create the GuardReasonerTrain dataset, which consists of 127K samples with 460K detailed reasoning steps. Then, we introduce reasoning SFT to unlock the reasoning capability of guard models. In addition, we present hard sample DPO to further strengthen their reasoning ability. In this manner, GuardReasoner achieves better performance, explainability, and generalizability. Extensive experiments and analyses on 13 benchmarks of 3 guardrail tasks demonstrate its superiority. Remarkably, GuardReasoner 8B surpasses GPT-4o+CoT by 5.74% and LLaMA Guard 3 8B by 20.84% F1 score on average. We release the training data, code, and models with different scales (1B, 3B, 8B) of GuardReasoner\footnote{https://github.com/yueliu1999/GuardReasoner/}.
this was a nice read, thx
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- Stepwise Reasoning Error Disruption Attack of LLMs (2024)
- Refining Input Guardrails: Enhancing LLM-as-a-Judge Efficiency Through Chain-of-Thought Fine-Tuning and Alignment (2025)
- From Drafts to Answers: Unlocking LLM Potential via Aggregation Fine-Tuning (2025)
- JailPO: A Novel Black-box Jailbreak Framework via Preference Optimization against Aligned LLMs (2024)
- Domaino1s: Guiding LLM Reasoning for Explainable Answers in High-Stakes Domains (2025)
- Token-Budget-Aware LLM Reasoning (2024)
- Layer-Level Self-Exposure and Patch: Affirmative Token Mitigation for Jailbreak Attack Defense (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 3
Datasets citing this paper 1
Spaces citing this paper 0
No Space linking this paper