Skip to content
AI Atlas

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

Published 18 Sept 2026arXiv:2609.19472

data quality89

Updated 4 h ago · first seen 18 Sept 2026

paper_01M2SEG2ZPK5MM8CAY9R4BRKWK

Abstract

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.

Authors

Authors 3

Alizishaan KhatriChiquita PrabhuOmkar Neogi

Linked names open researcher pages (created from the paper's author list; name-only, no affiliation unless a source states it). Unlinked names have no researcher record yet.

Organizations

Organizations 0

No organization stated. arXiv metadata does not carry affiliations; an organization is linked only when a model card or lab page cites the paper.

Models

Models introduced or described 0

Inbound described_by relations from model cards and documentation.

No model links this paper yet

Model pages link papers through their model cards and documentation; the relation is written only when a source states it.

Datasets

Datasets used 0

No dataset relation recorded.

Benchmarks

Benchmarks used 0

No benchmark relation recorded.

Code

Repositories & frameworks 0

No repository linked.

Timeline

Timeline 2

Full timeline →

Sources

Sources 2

Source documents
SourceDocumentTypeTierLast observedSnapshots
arXiv (Atom API + RSS)rss.arxiv.org/rss/cs.AI feedT1· Official4 h ago8
arXiv (Atom API + RSS)rss.arxiv.org/rss/cs.LG feedT1· Official4 h ago7

Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.