# When Truthful Representations Flip Under Deceptive Instructions?

arXiv: https://arxiv.org/abs/2507.22149 · HTML: https://mlsanity.com/paper/2507.22149/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in Probing deception, bias and beliefs, one of 13 directions in Mechanistic interpretability.
- Builds on 23 of its 38 references that are in the map.
- First posted on arXiv 29 Jul 2025.

## Placed in

- Mechanistic interpretability / [Probing deception, bias and beliefs](https://mlsanity.com/mechinterp/probing-deception-bias-and-beliefs/)

## Builds on (most cited first, 23 of 23)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Gemma 2: Improving Open Language Models at a Practical Size](https://mlsanity.com/paper/2408.00118/) (2024)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Wizard of Wikipedia: Knowledge-Powered Conversational agents](https://mlsanity.com/paper/1811.01241/) (2018)
- [Large Language Models can Strategically Deceive their Users when Put Under Pressure](https://mlsanity.com/paper/2311.07590/) (2023)
- [Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models](https://mlsanity.com/paper/2411.14257/) (2024)
- [Explore, Establish, Exploit: Red Teaming Language Models from Scratch](https://mlsanity.com/paper/2306.09442/) (2023)
- [How to Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questions](https://mlsanity.com/paper/2309.15840/) (2023)
- [Truth is Universal: Robust Detection of Lies in LLMs](https://mlsanity.com/paper/2407.12831/) (2024)
- [PURE: Turning Polysemantic Neurons Into Pure Features by Identifying Relevant Circuits](https://mlsanity.com/paper/2404.06453/) (2024)
- [Do LLMs "know" internally when they follow instructions?](https://mlsanity.com/paper/2410.14516/) (2024)
- [On the Universal Truthfulness Hyperplane Inside LLMs](https://mlsanity.com/paper/2407.08582/) (2024)
- [Understanding Linear Probing then Fine-tuning Language Models from NTK Perspective](https://mlsanity.com/paper/2405.16747/) (2024)
- [Sparse Distributed Memory is a Continual Learner](https://mlsanity.com/paper/2303.11934/) (2023)
- [FAEDKV: Infinite-Window Fourier Transform for Unbiased KV Cache Compression](https://mlsanity.com/paper/2507.20030/) (2025)
- [TruthEval: A Dataset to Evaluate LLM Truthfulness and Reliability](https://mlsanity.com/paper/2406.01855/) (2024)
- [Reimagining Linear Probing: Kolmogorov-Arnold Networks in Transfer Learning](https://mlsanity.com/paper/2409.07763/) (2024)
- [Interpretable Risk Mitigation in LLM Agent Systems](https://mlsanity.com/paper/2505.10670/) (2025)
- [Exploring the generalization of LLM truth directions on conversational formats](https://mlsanity.com/paper/2505.09807/) (2025)

## Built on by (most cited first, 10 of 10)

- [Quantized but Deceptive? A Multi-Dimensional Truthfulness Evaluation of Quantized LLMs](https://mlsanity.com/paper/2508.19432/) (2025)
- [The Truthfulness Spectrum Hypothesis](https://mlsanity.com/paper/2602.20273/) (2026)
- [Detecting Multi-Agent Collusion Through Multi-Agent Interpretability](https://mlsanity.com/paper/2604.01151/) (2026)
- [Pruning Weights but Not Truth: Safeguarding Truthfulness While Pruning LLMs](https://mlsanity.com/paper/2509.00096/) (2025)
- [Testing the Limits of Truth Directions in LLMs](https://mlsanity.com/paper/2604.03754/) (2026)
- [Decodability is Not Causality: Dissociating Probe Readouts from Behavioral Drivers via SAE Decomposition](https://mlsanity.com/paper/2609.18080/) (2026)
- [Diagnosing Correctness Probes under Self-Judgement Confounding](https://mlsanity.com/paper/2607.16799/) (2026)
- [VisuoAlign: Safety Alignment of LVLMs with Multimodal Tree Search](https://mlsanity.com/paper/2510.15948/) (2025)
- [Deep Minds and Shallow Probes](https://mlsanity.com/paper/2605.11448/) (2026)
- [Emergent Alignment](https://mlsanity.com/paper/2606.19527/) (2026)

## Abstract (arXiv)

Large language models (LLMs) tend to follow maliciously crafted instructions to generate deceptive responses, posing safety challenges. How deceptive instructions alter the internal representations of LLM compared to truthful ones remains poorly understood beyond output analysis. To bridge this gap, we investigate when and how these representations ``flip'', such as from truthful to deceptive, under deceptive versus truthful/neutral instructions. Analyzing the internal representations of Llama-3.1-8B-Instruct and Gemma-2-9B-Instruct on a factual verification task, we find the model's instructed True/False output is predictable via linear probes across all conditions based on the internal representation. Further, we use Sparse Autoencoders (SAEs) to show that the Deceptive instructions induce significant representational shifts compared to Truthful/Neutral representations (which are similar), concentrated in early-to-mid layers and detectable even on complex datasets. We also identify specific SAE features highly sensitive to deceptive instruction and use targeted visualizations to confirm distinct truthful/deceptive representational subspaces. % Our analysis pinpoints layer-wise and feature-level correlates of instructed dishonesty, offering insights for LLM detection and control. Our findings expose feature- and layer-level signatures of deception, offering new insights for detecting and mitigating instructed dishonesty in LLMs.