# Process or Result? Manipulated Ending Tokens Can Mislead Reasoning LLMs to Ignore the Correct Reasoning Steps

arXiv: https://arxiv.org/abs/2503.19326 · HTML: https://mlsanity.com/paper/2503.19326/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 20 of its 23 references that are in the map.
- First posted on arXiv 25 Mar 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)

## Builds on (most cited first, 20 of 20)

- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [Training Language Models to Self-Correct via Reinforcement Learning](https://mlsanity.com/paper/2409.12917/) (2024)
- [Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models](https://mlsanity.com/paper/2503.09567/) (2025)
- [From System 1 to System 2: A Survey of Reasoning Large Language Models](https://mlsanity.com/paper/2502.17419/) (2025)
- [Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View](https://mlsanity.com/paper/2310.02124/) (2023)
- [GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solvers](https://mlsanity.com/paper/2402.19255/) (2024)
- [LLMs Can Easily Learn to Reason from Demonstrations Structure, not content, is what matters!](https://mlsanity.com/paper/2502.07374/) (2025)
- [How well do Large Language Models perform in Arithmetic tasks?](https://mlsanity.com/paper/2304.02015/) (2023)
- [MathBench: Evaluating the Theory and Application Proficiency of LLMs with a Hierarchical Mathematics Benchmark](https://mlsanity.com/paper/2405.12209/) (2024)
- [The Hidden Risks of Large Reasoning Models: A Safety Assessment of R1](https://mlsanity.com/paper/2502.12659/) (2025)
- [Interpreting and Improving Large Language Models in Arithmetic Calculation](https://mlsanity.com/paper/2409.01659/) (2024)
- [Can Large Language Models Detect Errors in Long Chain-of-Thought Reasoning?](https://mlsanity.com/paper/2502.19361/) (2025)
- [Confidence v.s. Critique: A Decomposition of Self-Correction Capability for LLMs](https://mlsanity.com/paper/2412.19513/) (2024)
- [Language Models Do Hard Arithmetic Tasks Easily and Hardly Do Easy Arithmetic Tasks](https://mlsanity.com/paper/2406.02356/) (2024)
- [Mathematical Reasoning in Large Language Models: Assessing Logical and Arithmetic Errors across Wide Numerical Ranges](https://mlsanity.com/paper/2502.08680/) (2025)
- [Premise-Augmented Reasoning Chains Improve Error Identification in Math reasoning with LLMs](https://mlsanity.com/paper/2502.02362/) (2025)
- [BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning](https://mlsanity.com/paper/2501.03226/) (2025)
- [Learning From Correctness Without Prompting Makes LLM Efficient Reasoner](https://mlsanity.com/paper/2403.19094/) (2024)
- [ProgCo: Program Helps Self-Correction of Large Language Models](https://mlsanity.com/paper/2501.01264/) (2025)
- [FINEREASON: Evaluating and Improving LLMs' Deliberate Reasoning through Reflective Puzzle Solving](https://mlsanity.com/paper/2502.20238/) (2025)

## Built on by (most cited first, 9 of 9)

- [Safety in Large Reasoning Models: A Survey](https://mlsanity.com/paper/2504.17704/) (2025)
- [A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models](https://mlsanity.com/paper/2509.03871/) (2025)
- [MAD-Spear: A Conformity-Driven Prompt Injection Attack on Multi-Agent Debate Systems](https://mlsanity.com/paper/2507.13038/) (2025)
- [Practical Reasoning Interruption Attacks on Reasoning Large Language Models](https://mlsanity.com/paper/2505.06643/) (2025)
- [What Matters For Safety Alignment?](https://mlsanity.com/paper/2601.03868/) (2026)
- [Token-Efficient Prompt Injection Attack: Provoking Cessation in LLM Reasoning via Adaptive Token Compression](https://mlsanity.com/paper/2504.20493/) (2025)
- [Rethinking Reasoning: A Survey on Reasoning-based Backdoors in LLMs](https://mlsanity.com/paper/2510.07697/) (2025)
- [Red Teaming Large Reasoning Models](https://mlsanity.com/paper/2512.00412/) (2025)
- [From Perception to Reasoning: Deep Thinking Empowers Multimodal Large Language Models](https://mlsanity.com/paper/2511.12861/) (2025)

## Abstract (arXiv)

Recent reasoning large language models (LLMs) have demonstrated remarkable improvements in mathematical reasoning capabilities through long Chain-of-Thought. The reasoning tokens of these models enable self-correction within reasoning chains, enhancing robustness. This motivates our exploration: how vulnerable are reasoning LLMs to subtle errors in their input reasoning chains? We introduce "Compromising Thought" (CPT), a vulnerability where models presented with reasoning tokens containing manipulated calculation results tend to ignore correct reasoning steps and adopt incorrect results instead. Through systematic evaluation across multiple reasoning LLMs, we design three increasingly explicit prompting methods to measure CPT resistance, revealing that models struggle significantly to identify and correct these manipulations. Notably, contrary to existing research suggesting structural alterations affect model performance more than content modifications, we find that local ending token manipulations have greater impact on reasoning outcomes than structural changes. Moreover, we discover a security vulnerability in DeepSeek-R1 where tampered reasoning tokens can trigger complete reasoning cessation. Our work enhances understanding of reasoning robustness and highlights security considerations for reasoning-intensive applications.