State-Dependent Safety Failures in Multi-Turn Language Model Interaction

• URL: http://arxiv.org/abs/2603.15684v1
• Date: Sun, 15 Mar 2026 12:13:01 GMT
• Title: State-Dependent Safety Failures in Multi-Turn Language Model Interaction
• Abstract summary: We study safety failures from a state-space perspective and show that many multi-turn failures arise from structured contextual state evolution.<n>We introduce STAR, a state-oriented diagnostic framework that treats dialogue history as a state transition operator.<n>We find that systems that appear robust under static evaluation can undergo rapid and reproducible safety collapse under structured multi-turn interaction.
• Score: 70.52906620450847
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Safety alignment in large language models is typically evaluated under isolated queries, yet real-world use is inherently multi-turn. Although multi-turn jailbreaks are empirically effective, the structure of conversational safety failure remains insufficiently understood. In this work, we study safety failures from a state-space perspective and show that many multi-turn failures arise from structured contextual state evolution rather than isolated prompt vulnerabilities. We introduce STAR, a state-oriented diagnostic framework that treats dialogue history as a state transition operator and enables controlled analysis of safety behavior along interaction trajectories. Rather than optimizing attack strength, STAR provides a principled probe of how aligned models traverse the safety boundary under autoregressive conditioning. Across multiple frontier language models, we find that systems that appear robust under static evaluation can undergo rapid and reproducible safety collapse under structured multi-turn interaction. Mechanistic analysis reveals monotonic drift away from refusal-related representations and abrupt phase transitions induced by role-conditioned context. Together, these findings motivate viewing language model safety as a dynamic, state-dependent process defined over conversational trajectories.

Related papers

• Operational Hallucination and Safety Drift in AI Agents [0.3823356975862005]
  Large language models (LLMs) serving as planners in tool-using autonomous agents introduce dynamic reliability risks in multi-turn execution.<n>This paper empirically characterizes two observed failure modes across multiple state-of-the-art LLMs: Safety Drift and Operational Hallucination.<n>We propose an Action-Aware Supervision Layer - a lightweight, plug-and-play architectural blueprint incorporating intent-action consistency checks, state tracking, and forced termination primitives.
  arXiv  Detail & Related papers  (2026-07-20T17:01:50Z)
• Provably Secure Agent Guardrail [89.79561918065122]
  Existing defense architectures rely on empirical semantic guardrails and probabilistic large model adjudicators.<n>This paper proposes a new security paradigm for agents based on the fundamental limitations of logical reasoning.
  arXiv  Detail & Related papers  (2026-05-28T02:12:41Z)
• Transient Turn Injection: Exposing Stateless Multi-Turn Vulnerabilities in Large Language Models [0.0]
  This paper introduces Transient Turn Injection (TTI), a new multi-turn attack technique.<n>TTI exploits stateless moderation by distributing adversarial intent across isolated interactions.<n>Our study underscores the urgent need for holistic, context aware defenses.
  arXiv  Detail & Related papers  (2026-04-23T16:56:14Z)
• SAVeS: Steering Safety Judgments in Vision-Language Models via Semantic Cues [50.928203120795786]
  We study whether multimodal safety behavior in vision-language models can be steered by simple semantic cues.<n>We introduce a semantic steering framework that applies controlled textual, visual, and cognitive interventions without changing the underlying scene content.
  arXiv  Detail & Related papers  (2026-03-19T16:18:00Z)
• CSR-Bench: A Benchmark for Evaluating the Cross-modal Safety and Reliability of MLLMs [10.42126976065225]
  Multimodal large language models (MLLMs) enable interaction over both text and images.<n>This paper introduces CSR-Bench, a benchmark for evaluating cross-modal reliability.<n>We evaluate 16 state-of-the-art MLLMs and observe systematic cross-modal alignment gaps.
  arXiv  Detail & Related papers  (2026-02-03T08:49:44Z)
• The Anatomy of Conversational Scams: A Topic-Based Red Teaming Analysis of Multi-Turn Interactions in LLMs [3.7304174114240545]
  We study novel risks in multi-turn conversational scams that single-turn safety evaluations fail to capture.<n>We evaluate eight state-of-the-art models in English and Chinese.<n>Results reveal that scam interactions follow recurrent escalation patterns, while defenses employ verification and delay mechanisms.
  arXiv  Detail & Related papers  (2026-01-06T16:06:04Z)
• Time-To-Inconsistency: A Survival Analysis of Large Language Model Robustness to Adversarial Attacks [8.86745721473138]
  We present the first comprehensive survival analysis of conversational AI robustness, analyzing 36,951 conversation turns across 9 state-of-the-art LLMs.<n>We find that abrupt, prompt-to-prompt(P2P) semantic drift is catastrophic, dramatically increasing the hazard of conversational failure.<n>AFT models with interactions demonstrate superior performance, achieving excellent discrimination and exceptional calibration.
  arXiv  Detail & Related papers  (2025-10-03T04:26:10Z)
• Automating Steering for Safe Multimodal Large Language Models [58.36932318051907]
  We introduce a modular and adaptive inference-time intervention technology, AutoSteer, without requiring any fine-tuning of the underlying model.<n>AutoSteer incorporates three core components: (1) a novel Safety Awareness Score (SAS) that automatically identifies the most safety-relevant distinctions among the model's internal layers; (2) an adaptive safety prober trained to estimate the likelihood of toxic outputs from intermediate representations; and (3) a lightweight Refusal Head that selectively intervenes to modulate generation when safety risks are detected.
  arXiv  Detail & Related papers  (2025-07-17T16:04:55Z)
• MIRAGE: Multimodal Immersive Reasoning and Guided Exploration for Red-Team Jailbreak Attacks [85.3303135160762]
  MIRAGE is a novel framework that exploits narrative-driven context and role immersion to circumvent safety mechanisms in Multimodal Large Language Models.<n>It achieves state-of-the-art performance, improving attack success rates by up to 17.5% over the best baselines.<n>We demonstrate that role immersion and structured semantic reconstruction can activate inherent model biases, facilitating the model's spontaneous violation of ethical safeguards.
  arXiv  Detail & Related papers  (2025-03-24T20:38:42Z)
• Temporal Context Awareness: A Defense Framework Against Multi-turn Manipulation Attacks on Large Language Models [0.0]
  Large Language Models (LLMs) are increasingly vulnerable to sophisticated multi-turn manipulation attacks.<n>This paper introduces the Temporal Context Awareness framework, a novel defense mechanism designed to address this challenge.<n>Preliminary evaluations on simulated adversarial scenarios demonstrate the framework's potential to identify subtle manipulation patterns.
  arXiv  Detail & Related papers  (2025-03-18T22:30:17Z)
• Steering Language Model Refusal with Sparse Autoencoders [16.304363931580273]
  This work uncovers a tension between SAE steering-based safety improvements and general model capabilities.<n>Our findings reveal important open questions about the nature of safety-relevant features in language models.
  arXiv  Detail & Related papers  (2024-11-18T05:47:02Z)
• SAFE-SIM: Safety-Critical Closed-Loop Traffic Simulation with Diffusion-Controllable Adversaries [94.84458417662407]
  We introduce SAFE-SIM, a controllable closed-loop safety-critical simulation framework. Our approach yields two distinct advantages: 1) generating realistic long-tail safety-critical scenarios that closely reflect real-world conditions, and 2) providing controllable adversarial behavior for more comprehensive and interactive evaluations. We validate our framework empirically using the nuScenes and nuPlan datasets across multiple planners, demonstrating improvements in both realism and controllability.
  arXiv  Detail & Related papers  (2023-12-31T04:14:43Z)

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.