# The Struggle Between Continuation and Refusal: A Mechanistic Analysis of the Continuation-Triggered Jailbreak in LLMs

arXiv: https://arxiv.org/abs/2603.08234 · HTML: https://mlsanity.com/paper/2603.08234/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Jailbreaking, one of 12 directions in LLM safety and security.
- Builds on 30 of its 41 references that are in the map.
- First posted on arXiv 9 Mar 2026.

## Placed in

- LLM safety and security / [Jailbreaking](https://mlsanity.com/safety/jailbreaking/)
- Knowledge distillation / [Knowledge distillation for LLMs](https://mlsanity.com/distill/knowledge-distillation-for-llms/)

## Builds on (most cited first, 30 of 30)

- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Jailbreaking Black Box Large Language Models in Twenty Queries](https://mlsanity.com/paper/2310.08419/) (2023)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [In-context Learning and Induction Heads](https://mlsanity.com/paper/2209.11895/) (2022)
- [Contrastive Decoding: Open-ended Text Generation as Optimization](https://mlsanity.com/paper/2210.15097/) (2022)
- [JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models](https://mlsanity.com/paper/2404.01318/) (2024)
- [Large Language Models: A Survey](https://mlsanity.com/paper/2402.06196/) (2024)
- ["Do Anything Now": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Models](https://mlsanity.com/paper/2308.03825/) (2023)
- [Safety Alignment Should Be Made More Than Just a Few Tokens Deep](https://mlsanity.com/paper/2406.05946/) (2024)
- [Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation](https://mlsanity.com/paper/2310.06987/) (2023)
- [DeepInception: Hypnotize Large Language Model to Be Jailbreaker](https://mlsanity.com/paper/2311.03191/) (2023)
- [Towards Best Practices of Activation Patching in Language Models: Metrics and Methods](https://mlsanity.com/paper/2309.16042/) (2023)
- [Defending Against Indirect Prompt Injection Attacks With Spotlighting](https://mlsanity.com/paper/2403.14720/) (2024)
- [Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks](https://mlsanity.com/paper/2401.17263/) (2024)
- [The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections](https://mlsanity.com/paper/2510.09023/) (2025)
- [PromptArmor: Simple yet Effective Prompt Injection Defenses](https://mlsanity.com/paper/2507.15219/) (2025)
- [LLMs Encode Harmfulness and Refusal Separately](https://mlsanity.com/paper/2507.11878/) (2025)
- [Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety](https://mlsanity.com/paper/2502.05206/) (2025)
- [Interpreting and Improving Large Language Models in Arithmetic Calculation](https://mlsanity.com/paper/2409.01659/) (2024)
- [Activation Scaling for Steering and Interpreting Language Models](https://mlsanity.com/paper/2410.04962/) (2024)
- [Goal-guided Generative Prompt Injection Attack on Large Language Models](https://mlsanity.com/paper/2404.07234/) (2024)
- [Sugar-Coated Poison: Benign Generation Unlocks LLM Jailbreaking](https://mlsanity.com/paper/2504.05652/) (2025)
- [The Echo Chamber Multi-Turn LLM Jailbreak](https://mlsanity.com/paper/2601.05742/) (2026)
- [ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack](https://mlsanity.com/paper/2509.25843/) (2025)

## Built on by (most cited first, 1 of 1)

- [Jailbreak Context Lingers: Divergent Safety Routing and Its Cross-Task Predictability in Tool Agents](https://mlsanity.com/paper/2609.34686/) (2026)

## Abstract (arXiv)

With the rapid advancement of large language models (LLMs), the safety of LLMs has become a critical concern. Despite significant efforts in safety alignment, current LLMs remain vulnerable to jailbreaking attacks. However, the root causes of such vulnerabilities are still poorly understood, necessitating a rigorous investigation into jailbreak mechanisms across both academic and industrial communities. In this work, we focus on a continuation-triggered jailbreak phenomenon, whereby simply relocating a continuation-triggered instruction suffix can substantially increase jailbreak success rates. To uncover the intrinsic mechanisms of this phenomenon, we conduct a comprehensive mechanistic interpretability analysis at the level of attention heads. Through causal interventions and activation scaling, we show that this jailbreak behavior primarily arises from an inherent competition between the model's intrinsic continuation drive and the safety defenses acquired through alignment training. Furthermore, we perform a detailed behavioral analysis of the identified safety-critical attention heads, revealing notable differences in the behaviors of safety heads across different model architectures. Grounded in these mechanistic findings, we propose Head Competition Steering (HCS), a mechanistically grounded inference-time strategy that explicitly leverages the competition between safety heads and continuation heads to suppress harmful generation, and further distill its behavioral signal into a student model via knowledge distillation, achieving inference-time safety improvements without additional computational overhead.