When Prohibitions Become Permissions: Auditing Negation Sensitivity in Language Models

• URL: http://arxiv.org/abs/2601.21433v1
• Date: Thu, 29 Jan 2026 09:09:23 GMT
• Title: When Prohibitions Become Permissions: Auditing Negation Sensitivity in Language Models
• Abstract summary: Open-source models endorse prohibited actions 77% of the time under simple negation.<n>Financial scenarios prove twice as fragile as medical ones.<n>Findings point to a gap between what current alignment techniques achieve and what safe deployment requires.
• Score: 1.0671844383558033
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: When a user tells an AI system that someone "should not" take an action, the system ought to treat this as a prohibition. Yet many large language models do the opposite: they interpret negated instructions as affirmations. We audited 16 models across 14 ethical scenarios and found that open-source models endorse prohibited actions 77% of the time under simple negation and 100% under compound negation -- a 317% increase over affirmative framing. Commercial models fare better but still show swings of 19-128%. Agreement between models drops from 74% on affirmative prompts to 62% on negated ones, and financial scenarios prove twice as fragile as medical ones. These patterns hold under deterministic decoding, ruling out sampling noise. We present case studies showing how these failures play out in practice, propose the Negation Sensitivity Index (NSI) as a governance metric, and outline a tiered certification framework with domain-specific thresholds. The findings point to a gap between what current alignment techniques achieve and what safe deployment requires: models that cannot reliably distinguish "do X" from "do not X" should not be making autonomous decisions in high-stakes contexts.

Related papers

• Solving versus Verifying: Catching Contradictions in Tax Reasoning Systems [0.0]
  We study six recent models on SARA-derived tax cases under missing-fact and contradictory-fact perturbations.<n>We wire that verification call into a simple contradiction gate: one extra call that abstains when the model reports a conflict.
  arXiv  Detail & Related papers  (2026-09-05T06:38:02Z)
• The Sound of Absence: Audio-Language Embedding Models Struggle with Negation [58.04745279785462]
  We show that embedding models fail to encode negated sound concepts.<n>We introduce NegEval-Audio, a framework that converts datasets into two negation-aware tasks.
  arXiv  Detail & Related papers  (2026-07-14T02:55:00Z)
• RefusalBench: Why Refusal Rate Misranks Frontier LLMs on Biological Research Prompts [0.0]
  Strict refusal rates span 0.1% to 94.6% on identical prompts.<n>Nine of 18 frontier models exhibit a hedge-but-help partial-compliance pattern at dual-use tier that binary refusal metrics cannot detect.
  arXiv  Detail & Related papers  (2026-05-20T09:53:31Z)
• From Black Box to Glass Box: Cross-Model ASR Disagreement to Prioto Review in Ambient AI Scribe Documentation [43.148402136307716]
  Cross-model disagreement among heterogeneous ASR systems can act as a reference-free uncertainty signal.<n>We transcribed 50 publicly available medical audio clips with eight ASR systems spanning commercial APIs and open-source engines.<n>Low-agreement regions were enriched for content disagreements, with the content fraction increasing from 53.9% to 73.9% across quintiles of high-risk mass.
  arXiv  Detail & Related papers  (2026-03-02T13:02:13Z)
• A Coin Flip for Safety: LLM Judges Fail to Reliably Measure Adversarial Robustness [57.510025257780306]
  We show that existing validation protocols fail to account for substantial distribution shifts inherent to red-teaming.<n>We propose ReliableBench, a benchmark of behaviors that remain more consistently judgeable, and JudgeStressTest, a dataset designed to expose judge failures.
  arXiv  Detail & Related papers  (2026-02-04T15:13:35Z)
• Powerful Training-Free Membership Inference Against Autoregressive Language Models [3.9380576851378657]
  We present EZ-MIA, a membership inference attack that exploits a key observation: memorization manifests most strongly at error positions.<n>We introduce the Error Zone (EZ) score, which measures the directional imbalance of probability shifts at error positions.<n>Results establish that privacy risks of fine-tuned language models are substantially greater than previously understood.
  arXiv  Detail & Related papers  (2026-01-17T16:59:41Z)
• Are LLM Decisions Faithful to Verbal Confidence? [15.666596480779104]
  We introduce a framework designed to evaluate whether models adjust their abstention policies in response to varying error penalties.<n>Our evaluation of several frontier models reveals a critical dissociation: models are neither cost-aware when articulating their verbal confidence, nor strategically responsive when deciding whether to engage or abstain.<n>This indicates that calibrated verbal confidence scores may not be sufficient to create trustworthy and interpretable AI systems.
  arXiv  Detail & Related papers  (2026-01-12T17:49:51Z)
• Self-Transparency Failures in Expert-Persona LLMs: A Large-Scale Behavioral Audit [0.0]
  This study examines whether models exhibit self-transparency when assigned professional personas in high-stakes domains.<n>A Financial Advisor persona elicited 30.8% disclosure at the first prompt, while a Neurosurgeon persona elicited only 3.5%.<n>This creates the preconditions for a hypothesized Reverse Gell-Mann Amnesia effect, where appropriate disclosure in some domains leads users to overgeneralize trust to high-stakes contexts.
  arXiv  Detail & Related papers  (2025-11-26T16:41:49Z)
• Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning? [68.82210578851442]
  We investigate why safety alignment fails in reasoning models through a mechanistic interpretability lens.<n>Using a linear probing approach to trace refusal intentions across token positions, we discover a phenomenon termed as textbfrefusal cliff<n>We propose textbfCliff-as-a-Judge, a novel data selection method that identifies training examples exhibiting the largest refusal cliff to efficiently repair reasoning models' safety alignment.
  arXiv  Detail & Related papers  (2025-10-07T15:32:59Z)
• Measuring Language Model Hallucinations Through Distributional Correctness [7.106986689736826]
  A novel evaluation metric, the Distributional Correctness Score (DCS), is introduced to solve this problem.<n>DCS distinguishes between harmful overconfidence in wrong answers and uncertainty expressed through abstention, providing scores in an interpretable default range.<n>DCS is demonstrated to offer a more nuanced and aligned evaluation paradigm that incentivises models to express genuine uncertainty rather than guessing.
  arXiv  Detail & Related papers  (2025-10-05T17:50:42Z)
• Causal Understanding by LLMs: The Role of Uncertainty [43.87879175532034]
  Recent papers show LLMs achieve near-random accuracy in causal relation classification.<n>We investigate whether pretraining exposure to causal examples improves causal understanding.
  arXiv  Detail & Related papers  (2025-09-24T13:06:35Z)
• Confident, Calibrated, or Complicit: Probing the Trade-offs between Safety Alignment and Ideological Bias in Language Models in Detecting Hate Speech [0.916708284510944]
  We investigate the efficacy of Large Language Models (LLMs) in detecting implicit and explicit hate speech.<n>We find that censored models significantly outperform their uncensored counterparts in both accuracy and robustness.
  arXiv  Detail & Related papers  (2025-08-31T03:00:55Z)
• Do What? Teaching Vision-Language-Action Models to Reject the Impossible [53.40183895299108]
  Vision-Language-Action (VLA) models have demonstrated strong performance on a range of robotic tasks.<n>We propose Instruct-Verify-and-Act (IVA), a framework that detects when an instruction cannot be executed due to a false premise.<n>Our experiments show that IVA improves false premise detection accuracy by 97.56% over baselines.
  arXiv  Detail & Related papers  (2025-08-22T10:54:33Z)
• Mitigating Subgroup Disparities in Multi-Label Speech Emotion Recognition: A Pseudo-Labeling and Unsupervised Learning Approach [53.824673312331626]
  Implicit Demography Inference (IDI) module uses k-means clustering to mitigate bias in Speech Emotion Recognition (SER)<n>Experiments show that pseudo-labeling IDI reduces subgroup disparities, improving fairness metrics by over 28%.<n>Unsupervised IDI yields more than a 4.6% improvement in fairness metrics with a drop of less than 3.6% in SER performance.
  arXiv  Detail & Related papers  (2025-05-20T14:50:44Z)
• Exploring Response Uncertainty in MLLMs: An Empirical Evaluation under Misleading Scenarios [49.53589774730807]
  Multimodal large language models (MLLMs) have recently achieved state-of-the-art performance on tasks ranging from visual question answering to video understanding.<n>We reveal a response uncertainty phenomenon: twelve state-of-the-art open-source MLLMs overturn a previously correct answer in 65% of cases after receiving a single deceptive cue.
  arXiv  Detail & Related papers  (2024-11-05T01:11:28Z)
• Uncertainty in Language Models: Assessment through Rank-Calibration [65.10149293133846]
  Language Models (LMs) have shown promising performance in natural language generation. It is crucial to correctly quantify their uncertainty in responding to given inputs. We develop a novel and practical framework, termed $Rank$-$Calibration$, to assess uncertainty and confidence measures for LMs.
  arXiv  Detail & Related papers  (2024-04-04T02:31:05Z)
• Holistic Evaluation of Language Models [183.94891340168175]
  Language models (LMs) are becoming the foundation for almost all major language technologies, but their capabilities, limitations, and risks are not well understood. We present Holistic Evaluation of Language Models (HELM) to improve the transparency of language models.
  arXiv  Detail & Related papers  (2022-11-16T18:51:34Z)
• Provable Limitations of Acquiring Meaning from Ungrounded Form: What will Future Language Models Understand? [87.20342701232869]
  We investigate the abilities of ungrounded systems to acquire meaning. We study whether assertions enable a system to emulate representations preserving semantic relations like equivalence. We find that assertions enable semantic emulation if all expressions in the language are referentially transparent. However, if the language uses non-transparent patterns like variable binding, we show that emulation can become an uncomputable problem.
  arXiv  Detail & Related papers  (2021-04-22T01:00:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.