# Stable Reasoning, Unstable Responses: Mitigating LLM Deception via Stability Asymmetry

arXiv: https://arxiv.org/abs/2603.26846 · HTML: https://mlsanity.com/paper/2603.26846/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning.
- Builds on 30 of its 47 references that are in the map.
- First posted on arXiv 27 Mar 2026.

## Placed in

- LLM reasoning / [Prompting and chain-of-thought reasoning](https://mlsanity.com/lr/prompting-and-chain-of-thought-reasoning/)

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

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation](https://mlsanity.com/paper/2302.09664/) (2023)
- [The Internal State of an LLM Knows When It's Lying](https://mlsanity.com/paper/2304.13734/) (2023)
- [Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training](https://mlsanity.com/paper/2401.05566/) (2024)
- [Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation](https://mlsanity.com/paper/2503.11926/) (2025)
- [Measuring short-form factuality in large language models](https://mlsanity.com/paper/2411.04368/) (2024)
- [Alignment faking in large language models](https://mlsanity.com/paper/2412.14093/) (2024)
- [Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety](https://mlsanity.com/paper/2507.11473/) (2025)
- [Natural Emergent Misalignment from Reward Hacking in Production RL](https://mlsanity.com/paper/2511.18397/) (2025)
- [Auditing language models for hidden objectives](https://mlsanity.com/paper/2503.10965/) (2025)
- [The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems](https://mlsanity.com/paper/2503.03750/) (2025)
- [When Truth Is Overridden: Uncovering the Internal Origins of Sycophancy in Large Language Models](https://mlsanity.com/paper/2508.02087/) (2025)
- [The Traitors: Deception and Trust in Multi-Agent Language Model Simulations](https://mlsanity.com/paper/2505.12923/) (2025)
- [LMUnit: Fine-grained Evaluation with Natural Language Unit Tests](https://mlsanity.com/paper/2412.13091/) (2024)
- [Mitigating Deceptive Alignment via Self-Monitoring](https://mlsanity.com/paper/2505.18807/) (2025)
- [BeHonest: Benchmarking Honesty in Large Language Models](https://mlsanity.com/paper/2406.13261/) (2024)
- [Training LLMs for Honesty via Confessions](https://mlsanity.com/paper/2512.08093/) (2025)
- [When Thinking LLMs Lie: Unveiling the Strategic Deception in Representations of Reasoning Models](https://mlsanity.com/paper/2506.04909/) (2025)
- [Provably Efficient Generalized Lagrangian Policy Optimization for Safe Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2306.00212/) (2023)
- [Poser: Unmasking Alignment Faking LLMs by Manipulating Their Internals](https://mlsanity.com/paper/2405.05466/) (2024)
- [Can LLMs Lie? Investigation beyond Hallucination](https://mlsanity.com/paper/2509.03518/) (2025)
- [AI Deception: Risks, Dynamics, and Controls](https://mlsanity.com/paper/2511.22619/) (2025)
- [Alignment Faking - the Train -> Deploy Asymmetry: Through a Game-Theoretic Lens with Bayesian-Stackelberg Equilibria](https://mlsanity.com/paper/2511.17937/) (2025)
- [Think Before You Lie: How Reasoning Leads to Honesty](https://mlsanity.com/paper/2603.09957/) (2026)

## Abstract (arXiv)

As Large Language Models (LLMs) expand in capability and application scope, their trustworthiness becomes critical. A vital risk is intrinsic deception, wherein models strategically mislead users to achieve their own objectives. Existing alignment approaches based on chain-of-thought (CoT) monitoring supervise explicit reasoning traces. However, under optimization pressure, models are incentivized to conceal deceptive reasoning, rendering semantic supervision fundamentally unreliable. Grounded in cognitive psychology, we hypothesize that a deceptive LLM maintains a stable internal belief in its CoT while its external response remains fragile under perturbation. We term this phenomenon stability asymmetry and quantify it by measuring the contrast between internal CoT stability and external response stability under perturbation. Building on this structural signature, we propose the Stability Asymmetry Regularization (SAR), a novel alignment objective that penalizes this distributional asymmetry during reinforcement learning. Unlike CoT monitoring, SAR targets the statistical structure of model outputs, rendering it robust to semantic concealment. Extensive experiments confirm that stability asymmetry reliably identifies deceptive behavior, and that SAR effectively suppresses intrinsic deception without degrading general model capability.