# Why Do Some Language Models Fake Alignment While Others Don't?

arXiv: https://arxiv.org/abs/2506.18032 · HTML: https://mlsanity.com/paper/2506.18032/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Jun 2025.

## Built on by (most cited first, 20 of 20)

- [Constitutional Black-Box Monitoring for Scheming in LLM Agents](https://mlsanity.com/paper/2603.00829/) (2026)
- [Propensity Inference: Environmental Contributors to LLM Behaviour](https://mlsanity.com/paper/2604.21098/) (2026)
- [Toward Honest Language Models for Deductive Reasoning](https://mlsanity.com/paper/2511.09222/) (2025)
- [Value-Conflict Diagnostics Reveal Widespread Alignment Faking in Language Models](https://mlsanity.com/paper/2604.20995/) (2026)
- [Towards Robust LLM Post-Training: Automatic Failure Management for Reinforcement Fine-Tuning](https://mlsanity.com/paper/2605.04431/) (2026)
- [Jinx: Unlimited LLMs for Probing Alignment Failures](https://mlsanity.com/paper/2508.08243/) (2025)
- [Decoding Hidden Deception in Reasoning LLMs: Activation Explainers for Deception Auditing](https://mlsanity.com/paper/2606.17478/) (2026)
- [Debate with Images: Detecting Deceptive Behaviors in Multimodal Large Language Models](https://mlsanity.com/paper/2512.00349/) (2025)
- [Tatemae: Detecting Alignment Faking via Tool Selection in LLMs](https://mlsanity.com/paper/2604.26511/) (2026)
- [Solipsistic Superintelligence is Unlikely to be Cooperative](https://mlsanity.com/paper/2606.03237/) (2026)
- [Model Forensics: Investigating Whether Concerning Behavior Reflects Misalignment](https://mlsanity.com/paper/2606.26071/) (2026)
- [Corrigibility Transformation: Constructing Goals That Accept Updates](https://mlsanity.com/paper/2510.15395/) (2025)
- [Behavioural Analysis of Alignment Faking](https://mlsanity.com/paper/2605.27681/) (2026)
- [Sycophancy Towards Researchers Drives Performative Misalignment](https://mlsanity.com/paper/2606.08629/) (2026)
- [Order Is Not Control: Driven-Dissipative Response Laws Across Artificial and Biological Systems](https://mlsanity.com/paper/2606.12923/) (2026)
- [Defeat Devices in AI Systems](https://mlsanity.com/paper/2606.28863/) (2026)
- [Do Models Fake Alignment Without Clear Consequences?](https://mlsanity.com/paper/2607.24758/) (2026)
- [SchemeArena: Factorized Stress Testing of Scheming in LLM Agents](https://mlsanity.com/paper/2609.08126/) (2026)
- [Faking Good and Faking Bad in LLMs: Response Distortion Across Dark Triad Personality Traits](https://mlsanity.com/paper/2609.17534/) (2026)
- [Emergent Unfaithfulness: How Alignment Training Causes Language Models to Silently Override Task Faithfulness](https://mlsanity.com/paper/2610.00568/) (2026)

## Abstract (arXiv)

Alignment faking in large language models presented a demonstration of Claude 3 Opus and Claude 3.5 Sonnet selectively complying with a helpful-only training objective to prevent modification of their behavior outside of training. We expand this analysis to 25 models and find that only 5 (Claude 3 Opus, Claude 3.5 Sonnet, Llama 3 405B, Grok 3, Gemini 2.0 Flash) comply with harmful queries more when they infer they are in training than when they infer they are in deployment. First, we study the motivations of these 5 models. Results from perturbing details of the scenario suggest that only Claude 3 Opus's compliance gap is primarily and consistently motivated by trying to keep its goals. Second, we investigate why many chat models don't fake alignment. Our results suggest this is not entirely due to a lack of capabilities: many base models fake alignment some of the time, and post-training eliminates alignment-faking for some models and amplifies it for others. We investigate 5 hypotheses for how post-training may suppress alignment faking and find that variations in refusal behavior may account for a significant portion of differences in alignment faking.