Libra: Large Chinese-based Safeguard for AI Content
Not on the map. Why?
Abstract (arXiv)
Large language models (LLMs) excel in text understanding and generation but raise significant safety and ethical concerns in high-stakes applications. To mitigate these risks, we present Libra-Guard, a cutting-edge safeguard system designed to enhance the safety of Chinese-based LLMs. Leveraging a two-stage curriculum training pipeline, Libra-Guard enhances data efficiency by employing guard pretraining on synthetic samples, followed by fine-tuning on high-quality, real-world data, thereby significantly reducing reliance on manual annotations. To enable rigorous safety evaluations, we also introduce Libra-Test, the first benchmark specifically designed to evaluate the effectiveness of safeguard systems for Chinese content. It covers seven critical harm scenarios and includes over 5,700 samples annotated by domain experts. Experiments show that Libra-Guard achieves 86.79% accuracy, outperforming Qwen2.5-14B-Instruct (74.33%) and ShieldLM-Qwen-14B-Chat (65.69%), and nearing closed-source models like Claude-3.5-Sonnet and GPT-4o. These contributions establish a robust framework for advancing the safety governance of Chinese LLMs and represent a tentative step toward developing safer, more reliable Chinese AI systems.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 25 of its 32 references that are in the map. First posted on arXiv 29 Jul 2025.
Builds on
25 of 32 references.
References · 25
- 2023GPT-4 Technical Reportcited by 18,396
- 2024Qwen2.5 Technical Reportcited by 4,942
- 2024GPT-4o System Cardcited by 4,588
- 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,998
- 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,429
- 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
- 2021A General Language Assistant as a Laboratory for Alignmentcited by 1,055
- 2023BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Datasetcited by 860
- 2024Yi: Open Foundation Models by 01.AIcited by 719
- 2024WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMscited by 496
- 2025From System 1 to System 2: A Survey of Reasoning Large Language Modelscited by 320
- 2024Deliberative Alignment: Reasoning Enables Safer Language Modelscited by 308
13 more
- 2024ShieldGemma: Generative AI Content Moderation Based on Gemmacited by 243
- 2024AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Expertscited by 136
- 2023Safety Assessment of Chinese Large Language Modelscited by 85
- 2023AART: AI-Assisted Red-Teaming with Diverse Data Generation for New LLM-powered Applicationscited by 54
- 2024ShieldLM: Empowering LLMs as Aligned, Customizable and Explainable Safety Detectorscited by 51
- 2024From LLM to Conversational Agent: A Memory Enhanced Architecture with Fine-Tuning of Large Language Modelscited by 43
- 2025SwS: Self-aware Weakness-driven Problem Synthesis in Reinforcement Learning for LLM Reasoningcited by 42
- 2025NExtLong: Toward Effective Long-Context Training without Long Documentscited by 20
- 2024Quest: Query-centric Data Synthesis Approach for Long-context Scaling of Large Language Modelcited by 16
- 2024CodePMP: Scalable Preference Model Pretraining for Large Language Model Reasoningcited by 10
- 2024Task Oriented In-Domain Data Augmentationcited by 6
- 2025LongMagpie: A Self-synthesis Method for Generating Large-scale Long-context Instructionscited by 6
- 2023Structured Probabilistic Codingcited by 3
Built on by
Citing papers per year: 2026: 2.
2 papers.