Prefilling the Reasoning Channel: Output-Prefix Attacks on Reasoning LLMs

arXiv:2609.29775 · cs.CR, cs.AI · Submitted 2026-09-24 · Read on arXiv

cs.CR, cs.AI

Submitted: 2026-09-24

Updated: 2026-09-24

Code: https://github.com/lukasbruna/output-prefix-attack

Terminology

• Alignment — 7× in this paper · explained in 1 episode
• Latent Reasoning — 7× in this paper · explained in 2 episodes
• Reinforcement Learning — 4× in this paper · explained in 4 episodes
• Adam — 3× in this paper · explained in 2 episodes
• Attack Success Rate — 3× in this paper
• Chain-of-Thought (CoT) — 2× in this paper · explained in 4 episodes
• Chain-of-Thought (CoT) Prompting — 2× in this paper · explained in 3 episodes
• Chain-of-Thought Prompting — 2× in this paper · explained in 2 episodes
• Core — 2× in this paper · explained in 2 episodes
• Fine-Tuning — 2× in this paper · explained in 5 episodes
• LLM judge — 2× in this paper · explained in 2 episodes
• Robustness — 2× in this paper · explained in 5 episodes
• Transformer — 2× in this paper · explained in 2 episodes
• Addressability — 1× in this paper · explained in 1 episode
• Adversarial Attacks — 1× in this paper · explained in 1 episode
• Attack Success Rate (ASR) — 1× in this paper · explained in 2 episodes
• Benchmark — 1× in this paper · explained in 2 episodes
• Explainability — 1× in this paper · explained in 3 episodes
• Generative Model — 1× in this paper · explained in 2 episodes
• Generative Models — 1× in this paper · explained in 2 episodes
• Hallucination — 1× in this paper · explained in 3 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Latent Space — 1× in this paper · explained in 7 episodes

Sources

• Constitutional AI: Harmlessness from AI Feedback
• ChatInject: Abusing Chat Templates for Prompt Injection in LLM Agents
• Training Large Language Models to Reason in a Continuous Latent Space
• OverThink: Slowdown Attacks on Reasoning LLMs
• Reinforcement Learning from Human Feedback
• Show Your Work: Scratchpads for Intermediate Computation with Language Models
• Prefill Awareness in Large Language Models
• Universal and Transferable Adversarial Attacks on Aligned Language Models

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs