{"ID":2880501,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.14128","arxiv_id":"2508.14128","title":"CCFC: Core \u0026 Core-Full-Core Dual-Track Defense for LLM Jailbreak Protection","abstract":"Jailbreak attacks pose a serious challenge to the safe deployment of large language models (LLMs). We introduce CCFC (Core \u0026 Core-Full-Core), a dual-track, prompt-level defense framework designed to mitigate LLMs' vulnerabilities from prompt injection and structure-aware jailbreak attacks. CCFC operates by first isolating the semantic core of a user query via few-shot prompting, and then evaluating the query using two complementary tracks: a core-only track to ignore adversarial distractions (e.g., toxic suffixes or prefix injections), and a core-full-core (CFC) track to disrupt the structural patterns exploited by gradient-based or edit-based attacks. The final response is selected based on a safety consistency check across both tracks, ensuring robustness without compromising on response quality. We demonstrate that CCFC cuts attack success rates by 50-75% versus state-of-the-art defenses against strong adversaries (e.g., DeepInception, GCG), without sacrificing fidelity on benign queries. Our method consistently outperforms state-of-the-art prompt-level defenses, offering a practical and effective solution for safer LLM deployment.","short_abstract":"Jailbreak attacks pose a serious challenge to the safe deployment of large language models (LLMs). We introduce CCFC (Core \u0026 Core-Full-Core), a dual-track, prompt-level defense framework designed to mitigate LLMs' vulnerabilities from prompt injection and structure-aware jailbreak attacks. CCFC operates by first isolat...","url_abs":"https://arxiv.org/abs/2508.14128","url_pdf":"https://arxiv.org/pdf/2508.14128v1","authors":"[\"Jiaming Hu\",\"Haoyu Wang\",\"Debarghya Mukherjee\",\"Ioannis Ch. Paschalidis\"]","published":"2025-08-19T04:17:21Z","proceeding":"cs.CR","tasks":"[\"cs.CR\",\"cs.AI\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}