# Output Length Effect on DeepSeek-R1's Safety in Forced Thinking

arXiv: https://arxiv.org/abs/2503.01923 · HTML: https://mlsanity.com/paper/2503.01923/

## Key facts

- Cited by 6 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 14 of its 26 references that are in the map.
- First posted on arXiv 2 Mar 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 14 of 14)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [The Impact of Reasoning Step Length on Large Language Models](https://mlsanity.com/paper/2401.04925/) (2024)
- [Same Task, More Tokens: the Impact of Input Length on the Reasoning Performance of Large Language Models](https://mlsanity.com/paper/2402.14848/) (2024)
- [Concise Thoughts: Impact of Output Length on LLM Reasoning and Cost](https://mlsanity.com/paper/2407.19825/) (2024)
- [Let Me Speak Freely? A Study on the Impact of Format Restrictions on Performance of Large Language Models](https://mlsanity.com/paper/2408.02442/) (2024)
- [Towards Foundation Models for Knowledge Graph Reasoning](https://mlsanity.com/paper/2310.04562/) (2023)
- [A Queueing Theoretic Perspective on Low-Latency LLM Inference with Variable Token Length](https://mlsanity.com/paper/2407.05347/) (2024)
- [DrJAX: Scalable and Differentiable MapReduce Primitives in JAX](https://mlsanity.com/paper/2403.07128/) (2024)
- [Recurrent Memory for Online Interdomain Gaussian Processes](https://mlsanity.com/paper/2502.08736/) (2025)
- [Distributed Monitoring for Data Distribution Shifts in Edge-ML Fraud Detection](https://mlsanity.com/paper/2401.05219/) (2024)
- [Camouflage is all you need: Evaluating and Enhancing Language Model Robustness Against Camouflage Adversarial Attacks](https://mlsanity.com/paper/2402.09874/) (2024)
- [SwinFi: a CSI Compression Method based on Swin Transformer for Wi-Fi Sensing](https://mlsanity.com/paper/2405.03957/) (2024)
- [Utilitarian Distortion with Predictions](https://mlsanity.com/paper/2502.06489/) (2025)

## Built on by (most cited first, 6 of 6)

- [Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models](https://mlsanity.com/paper/2503.16419/) (2025)
- [Efficient Reasoning Models: A Survey](https://mlsanity.com/paper/2504.10903/) (2025)
- [100 Days After DeepSeek-R1: A Survey on Replication Studies and More Directions for Reasoning Language Models](https://mlsanity.com/paper/2505.00551/) (2025)
- [A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models](https://mlsanity.com/paper/2509.03871/) (2025)
- [Trade-offs in Large Reasoning Models: An Empirical Analysis of Deliberative and Adaptive Reasoning over Foundational Capabilities](https://mlsanity.com/paper/2503.17979/) (2025)
- [Towards Concise and Adaptive Thinking in Large Reasoning Models: A Survey](https://mlsanity.com/paper/2507.09662/) (2025)

## Abstract (arXiv)

Large Language Models (LLMs) have demonstrated strong reasoning capabilities, but their safety under adversarial conditions remains a challenge. This study examines the impact of output length on the robustness of DeepSeek-R1, particularly in Forced Thinking scenarios. We analyze responses across various adversarial prompts and find that while longer outputs can improve safety through self-correction, certain attack types exploit extended generations. Our findings suggest that output length should be dynamically controlled to balance reasoning effectiveness and security. We propose reinforcement learning-based policy adjustments and adaptive token length regulation to enhance LLM safety.