DeepKnown-Guard: A Proprietary Model-Based Safety Response Framework for AI Agents

TL;DR

This paper introduces DeepKnown-Guard, a safety framework for LLMs that enhances risk detection and response at input and output levels, achieving high safety scores and robustness in critical applications.

Contribution

The paper presents a novel, proprietary safety response framework combining fine-grained risk classification and retrieval-augmented generation to improve LLM safety and trustworthiness.

Findings

Achieved 99.3% risk recall rate in input safety classification.

Attained 100% safety score on high-risk test set.

Significantly outperformed baseline safety models on public benchmarks.

Abstract

With the widespread application of Large Language Models (LLMs), their associated security issues have become increasingly prominent, severely constraining their trustworthy deployment in critical domains. This paper proposes a novel safety response framework designed to systematically safeguard LLMs at both the input and output levels. At the input level, the framework employs a supervised fine-tuning-based safety classification model. Through a fine-grained four-tier taxonomy (Safe, Unsafe, Conditionally Safe, Focused Attention), it performs precise risk identification and differentiated handling of user queries, significantly enhancing risk coverage and business scenario adaptability, and achieving a risk recall rate of 99.3%. At the output level, the framework integrates Retrieval-Augmented Generation (RAG) with a specifically fine-tuned interpretation model, ensuring all responses are grounded in a real-time, trustworthy knowledge base. This approach eliminates information fabrication and enables result traceability. Experimental results demonstrate that our proposed safety control model achieves a significantly higher safety score on public safety evaluation benchmarks compared to the baseline model, TinyR1-Safety-8B. Furthermore, on our proprietary high-risk test set, the framework's components attained a perfect 100% safety score, validating their exceptional protective capabilities in complex risk scenarios. This research provides an effective engineering pathway for building high-security, high-trust LLM applications.