# CHILLGuard: Towards Fine-Grained Chinese LLM Safety Guardrail with Scalable Data Construction and Model-aware Preference Alignment

arXiv: https://arxiv.org/abs/2606.15396 · HTML: https://mlsanity.com/paper/2606.15396/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Safety alignment and jailbreak attacks, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 22 of its 48 references that are in the map.
- First posted on arXiv 13 Jun 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Safety alignment and jailbreak attacks](https://mlsanity.com/alignment/safety-alignment-and-jailbreak-attacks/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 22 of 22)

- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation](https://mlsanity.com/paper/2402.03216/) (2024)
- [BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset](https://mlsanity.com/paper/2307.04657/) (2023)
- [ORPO: Monolithic Preference Optimization without Reference Model](https://mlsanity.com/paper/2403.07691/) (2024)
- [WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs](https://mlsanity.com/paper/2406.18495/) (2024)
- [NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails](https://mlsanity.com/paper/2310.10501/) (2023)
- [ToxicChat: Unveiling Hidden Challenges of Toxicity Detection in Real-World User-AI Conversation](https://mlsanity.com/paper/2310.17389/) (2023)
- [Attacks, Defenses and Evaluations for LLM Conversation Safety: A Survey](https://mlsanity.com/paper/2402.09283/) (2024)
- [PolyGuard: A Multilingual Safety Moderation Tool for 17 Languages](https://mlsanity.com/paper/2504.04377/) (2025)
- [GIFD: A Generative Gradient Inversion Method with Feature Domain Optimization](https://mlsanity.com/paper/2308.04699/) (2023)
- [One Perturbation is Enough: On Generating Universal Adversarial Perturbations against Vision-Language Pre-training Models](https://mlsanity.com/paper/2406.05491/) (2024)
- [An Empirical Study on Challenges for LLM Application Developers](https://mlsanity.com/paper/2408.05002/) (2024)
- [DAMA: Data- and Model-aware Alignment of Multi-modal LLMs](https://mlsanity.com/paper/2502.01943/) (2025)
- [Revisiting Backdoor Attacks on LLMs: A Stealthy and Practical Poisoning Framework via Harmless Inputs](https://mlsanity.com/paper/2505.17601/) (2025)
- [Retrievals Can Be Detrimental: Unveiling the Backdoor Vulnerability of Retrieval-Augmented Diffusion Models](https://mlsanity.com/paper/2501.13340/) (2025)
- [JPRO: Automated Multimodal Jailbreaking via Multi-Agent Collaboration Framework](https://mlsanity.com/paper/2511.07315/) (2025)
- [Editable-DeepSC: Reliable Cross-Modal Semantic Communications for Facial Editing](https://mlsanity.com/paper/2411.15702/) (2024)
- [Enhancing Gradient Inversion Attacks in Federated Learning via Hierarchical Feature Optimization](https://mlsanity.com/paper/2604.00955/) (2026)
- [Reasoning Matters: Mitigate Hallucination in Multimodal Large Reasoning Models via Reasoning-Conditioned Preference Optimization](https://mlsanity.com/paper/2605.27906/) (2026)
- [Bypassing Copyright Protection in Diffusion-based Customization via Two-Stage Latent Feature Optimization](https://mlsanity.com/paper/2606.09909/) (2026)

## Abstract (arXiv)

Malicious content generated from large language models (LLMs) could pose severe safety risks and ethical concerns. While existing LLM safety guardrails excel in English or multilingual settings, they lack adaptation to Chinese-specific regulatory policies, cultural context, and linguistic nuances, failing to support fine-grained risk classification for diverse deployment needs. In this paper, we introduce a 5-macro, 31-micro category fine-grained risk taxonomy for Chinese scenarios, and build CHILLGuard: a dedicated Chinese LLM content safety guardrail. To address the critical scarcity of high-quality annotated Chinese safety data, we propose a scalable multi-stage data construction pipeline: we expand multi-source corpus via retrieval-augmented generation, generate implicit harmful samples through prompt engineering rewriting, and refine high-quality data via multi-model voting-based label calibration. Based on this, we build CHILLGuardTrain, a large-scale training set with 405,007 samples, and CHILLGuardTest, a rigorously curated annotated test set with 51,745 samples. We then train CHILLGuard on CHILLGuardTrain under a generator-classifier collaborative framework via Model-aware Direct Preference Optimization. Extensive experiments under multiple settings demonstrate the state-of-the-art performance of CHILLGuard, e.g., a 15.92% relative improvement of F1 score over Qwen3Guard-8B-Strict on our benchmark. We release our resources at https://github.com/cswbyu/CHILLGuard.