# Safety Beyond the Interface: Detecting Harm via Latent States in Large Language Models

arXiv: https://arxiv.org/abs/2609.19472 · HTML: https://mlsanity.com/paper/2609.19472/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 29 of its 33 references that are in the map.
- First posted on arXiv 16 Sep 2026.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 29 of 29)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset](https://mlsanity.com/paper/2307.04657/) (2023)
- [JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models](https://mlsanity.com/paper/2404.01318/) (2024)
- [WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs](https://mlsanity.com/paper/2406.18495/) (2024)
- [Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks](https://mlsanity.com/paper/2404.02151/) (2024)
- [Alignment faking in large language models](https://mlsanity.com/paper/2412.14093/) (2024)
- [ShieldGemma: Generative AI Content Moderation Based on Gemma](https://mlsanity.com/paper/2407.21772/) (2024)
- [Jailbreaking Attack against Multimodal Large Language Model](https://mlsanity.com/paper/2402.02309/) (2024)
- [Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails](https://mlsanity.com/paper/2501.09004/) (2025)
- [Attention Tracker: Detecting Prompt Injection Attacks in LLMs](https://mlsanity.com/paper/2411.00348/) (2024)
- [Building Guardrails for Large Language Models](https://mlsanity.com/paper/2402.01822/) (2024)
- [Towards Understanding Safety Alignment: A Mechanistic Perspective from Safety Neurons](https://mlsanity.com/paper/2406.14144/) (2024)
- [Constitutional Classifiers++: Efficient Production-Grade Defenses against Universal Jailbreaks](https://mlsanity.com/paper/2601.04603/) (2026)
- [Llama Guard 3-1B-INT4: Compact and Efficient Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2411.17713/) (2024)
- [Bypassing LLM Guardrails: An Empirical Analysis of Evasion Attacks against Prompt Injection and Jailbreak Detection Systems](https://mlsanity.com/paper/2504.11168/) (2025)
- [Large Language Models Encode Semantics and Alignment in Linearly Separable Representations](https://mlsanity.com/paper/2507.09709/) (2025)
- [LLMScan: Causal Scan for LLM Misbehavior Detection](https://mlsanity.com/paper/2410.16638/) (2024)
- [Guiding not Forcing: Enhancing the Transferability of Jailbreaking Attacks on LLMs via Removing Superfluous Constraints](https://mlsanity.com/paper/2503.01865/) (2025)
- [Next-Generation LLM for UAV: From Natural Language to Autonomous Flight](https://mlsanity.com/paper/2510.21739/) (2025)
- [Probing Latent Subspaces in LLM for AI Security: Identifying and Manipulating Adversarial States](https://mlsanity.com/paper/2503.09066/) (2025)
- [Lightweight Safety Guardrails Using Fine-tuned BERT Embeddings](https://mlsanity.com/paper/2411.14398/) (2024)
- [AutoAdv: Automated Adversarial Prompting for Multi-Turn Jailbreaking of Large Language Models](https://mlsanity.com/paper/2507.01020/) (2025)
- [DeepContext: Stateful Real-Time Detection of Multi-Turn Adversarial Intent Drift in LLMs](https://mlsanity.com/paper/2602.16935/) (2026)
- [Bypassing Prompt Guards in Production with Controlled-Release Prompting](https://mlsanity.com/paper/2510.01529/) (2025)
- [Hiding in Plain Sight: A Steganographic Approach to Stealthy LLM Jailbreaks](https://mlsanity.com/paper/2505.16765/) (2025)

## Built on by (most cited first, 1 of 1)

- [Same Probe, Different Numbers: Are Activation Probes Robust to Inference-Time Numerical Non-Determinism?](https://mlsanity.com/paper/2609.31796/) (2026)

## Abstract (arXiv)

Autonomous systems increasingly rely on Large Language Models (LLMs) yet the safety infrastructure surrounding these models introduces latency and compute overhead. This limits utility in resource-constrained, time-critical deployments. Existing external guardrail models remain blind to the model's internal workings, creating a fundamental assurance gap. We ask: does the model already know when the content is harmful? We extract activations from LLaMA-3.1-8B and train lightweight MLP classifier probes (12.6M parameters) to detect harmful prompts. Evaluated on WildJailbreak, Beavertails, and AEGIS 2.0, our probes achieve F1 scores of 99%, 83%, and 84%, respectively competitive with 1000x larger guard models while cutting latency and compute costs.