Polarity Blindness in LLM Safety: A Systematic Study of Negation, Antonym Substitution, and Adversarial Framing
Factual evidence
What the source reports
Researchers identify 'polarity blindness' in LLMs, showing safety guardrails fail when harmful prompts use negation or antonyms.
Inspect the evidence
- Inclusion basis
- Enterprise AI
- Publisher and source type
- arXiv cs.CL — Computation and Language · RESEARCH
- Published by source
- 21 October 2026
- Collected by OneBench
- 22 Oct 2026, 03:01 UK
Stored source excerpt
arXiv:2610.27908v1 Announce Type: new Abstract: Negation remains a persistent challenge for language models across NLP tasks and modalities. We show that this weakness can undermine…
Short excerpt from the collected text, not the full source. Use the source link to read it in context.
The factual summary is a OneBench synthesis, not a quotation or independent verification. Collection time is not publication time. Open the source for its full context; related reporting can share the same underlying announcement.
OneBench interpretation
Institutional assessment
So what
Standard red-teaming misses simple syntactic evasions, exposing LLM guardrails to trivial jailbreaks across automated workflows.
Do what
Review LLM guardrail test suites with risk teams to ensure adversarial evaluations include negation and polarity-inversion test cases.