Research map / Large language models / Retrieval and knowledge
Rescue: Retrieval Augmented Secure Code Generation
Jiahao Shi, Tianyi Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| construct security knowledge base instead of using raw security data (ablation) | Rescue (DeepSeek-Coder-V2-Lite) · CodeGuard+ | SecurePass@1: 55.6 -> 65.6 | Improved |
| construct security knowledge base instead of using raw security data (ablation) | Rescue (Qwen2.5-Coder-7B) · CodeGuard+ | SecurePass@1: 53.9 -> 64.8 | Improved |
| construct security knowledge base instead of using raw security data (ablation) | Rescue (Llama3.1-8B) · CodeGuard+ | SecurePass@1: 45.3 -> 56.2 | Improved |
| add security guidelines to retrieved examples (ablation) | Rescue (DeepSeek-Coder-V2-Lite) · CodeGuard+ | SecurePass@1: 63.3 -> 65.6 | Improved |
| add security guidelines to retrieved examples (ablation) | Rescue (Qwen2.5-Coder-7B) · CodeGuard+ | SecurePass@1: 63.9 -> 64.8 | Improved |
| add security guidelines to retrieved examples (ablation) | Rescue (Llama3.1-8B) · CodeGuard+ | SecurePass@1: 51.4 -> 56.2 | Improved |
| add program slicing (ablation) | Rescue (DeepSeek-Coder-V2-Lite) · CodeGuard+ | SecurePass@1: 61.0 -> 65.6 | Improved |
| add program slicing (ablation) | Rescue (Qwen2.5-Coder-7B) · CodeGuard+ | SecurePass@1: 62.6 -> 64.8 | Improved |
| add program slicing (ablation) | Rescue (Llama3.1-8B) · CodeGuard+ | SecurePass@1: 52.7 -> 56.2 | Improved |
| slice code examples in the generation stage (ablation) | Rescue (DeepSeek-Coder-V2-Lite) · CodeGuard+ | input tokens: 753 -> 503 | Improved |
| add vulnerability cause analysis facet to API and code facets (ablation) | Rescue (DeepSeek-Coder-V2-Lite) · CodeGuard+ | SecurePass@1: 62.7 -> 65.6 | Improved |
| add API pattern and vulnerability cause facets to code similarity (ablation) | Rescue (DeepSeek-Coder-V2-Lite) · CodeGuard+ | SecurePass@1: 57.1 -> 65.6 | Improved |
Closest papers by meaning
- Repoformer: Selective Retrieval for Repository-Level Code Completion (ICML 2024 · similarity 0.87)
- Follow My Instruction and Spill the Beans: Scalable Data Extraction from Retrieval-Augmented Generation Systems (ICLR 2025 · similarity 0.87)
- ProSec: Fortifying Code LLMs with Proactive Security Alignment (ICML 2025 · similarity 0.85)
- A Theory for Token-Level Harmonization in Retrieval-Augmented Generation (ICLR 2025 · similarity 0.85)
- PinNet: Pinpoint Instructive Information for Retrieval Augmented Code-to-Text Generation (ICML 2024 · similarity 0.85)
- SecCodePRM: A Process Reward Model for Code Security (ICML 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
