AI Threat Alert
TrailBlazer: History-Guided Reinforcement Learning for Black-Box LLM Jailbreaking
historical vulnerability signals in reinforcement learning-driven jailbreak strategies and offer a principled pathway for advancing adversarial research on LLM safeguards
Jailbreaking Attacks vs. Content Safety Filters: How Far Are We in the LLM Safety Arms Race?
moderation filters. To address this gap, we present the first systematic evaluation of jailbreak attacks targeting LLM safety alignment, assessing their success across the full inference pipeline, including both input
LITMUS: Benchmarking Behavioral Jailbreaks of LLM Agents in Real OS Environments
rapid proliferation of LLM-based autonomous agents in real operating system environments introduces a new category of safety risk beyond content safety: behavior jailbreak, where an adversary induces an agent
Active Honeypot Guardrail System: Probing and Confirming Multi-Turn LLM Jailbreaks
Large language models (LLMs) are increasingly vulnerable to multi-turn jailbreak attacks, where adversaries iteratively elicit harmful behaviors that bypass single-turn safety filters. Existing defenses predominantly rely on passive
SpatialJB: How Text Distribution Art Becomes the "Jailbreak Key" for LLM Guardrails
powerful capabilities, they remain vulnerable to jailbreak attacks, which is a critical barrier to their safe web real-time application. Current commercial LLM providers deploy output guardrails to filter harmful
How Reliable Is Your Jailbreak Judge? Calibration and Adversarial Robustness of Automated ASR Scoring
Almost every paper on LLM jailbreaks and prompt injection reports an attack-success rate (ASR), and that number is assigned not by people but by an automated judge: either
TRYLOCK: Defense-in-Depth Against LLM Jailbreaks via Layered Preference and Representation Engineering
Large language models remain vulnerable to jailbreak attacks, and single
NEXUS: Network Exploration for eXploiting Unsafe Sequences in Multi-Turn LLM Jailbreaks
Language Models (LLMs) have revolutionized natural language processing but remain vulnerable to jailbreak attacks, especially multi-turn jailbreaks that distribute malicious intent across benign exchanges and bypass alignment mechanisms. Existing
Provable Defense Framework for LLM Jailbreaks via Noise-Augumented Alignment
Large Language Models (LLMs) remain vulnerable to adaptive jailbreaks that
RedTWIZ: Diverse LLM Red Teaming via Adaptive Attack Planning
driven by three major research streams: (1) robust and systematic assessment of LLM conversational jailbreaks; (2) a diverse generative multi-turn attack suite, supporting compositional, realistic and goal-oriented jailbreak
The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections
How should we evaluate the robustness of language model defenses
ALERT: Zero-shot LLM Jailbreak Detection via Internal Discrepancy Amplification
Despite rich safety alignment strategies, large language models (LLMs) remain
Causal Front-Door Adjustment for Robust Jailbreak Attacks on LLMs
causal perspective. Then, we propose the Causal Front-Door Adjustment Attack (CFA{$^2$}) to jailbreak LLM, which is a framework that leverages Pearl's Front-Door Criterion to sever
Mechanistic Interpretability of LLM Jailbreaks via Internal Attribution Graphs
language models (LLMs) exhibit remarkable capabilities but remain highly vulnerable to adversarial prompts and jailbreak attacks. Existing approaches primarily analyze these failures through input-output behaviors or attribution methods, offering
Sentra-Guard: A Multilingual Human-AI Framework for Real-Time Defense Against Adversarial LLM Jailbreaks
This paper presents a real-time modular defense system named
Formalization Driven LLM Prompt Jailbreaking via Reinforcement Learning
challenges. For instance, prompt jailbreaking attacks involve adversaries crafting sophisticated prompts to elicit responses from LLMs that deviate from human values. To uncover vulnerabilities in LLM alignment methods, we propose
MJ: Multi-turn LLM Jailbreaking via Decomposed Credit Assignment
Modern large language models (LLMs) operate in interactive multi-turn
MEEA: Mere Exposure Effect-Driven Confrontational Optimization for LLM Jailbreaking
optimizes them using a simulated annealing strategy guided by semantic similarity, toxicity, and jailbreak effectiveness. Extensive experiments on both closed-source and open-source models, including GPT-4, Claude
ALMGuard: Safety Shortcuts and Where to Find Them as Guardrails for Audio-Language Models
defenses directly transferred from traditional audio adversarial attacks or text-based Large Language Model (LLM) jailbreaks are largely ineffective against these ALM-specific threats. To address this issue, we propose
SoK: Intent-Oriented Systematization of Multi-Turn LLM Jailbreaks
Large Language Models (LLMs) are increasingly deployed in interactive settings