Exploring Jailbreak Vulnerability Across Depth-Based Dynamic Inference in Large Language Models

cs.LG, cs.AI, cs.CL, cs.CR

Submitted: 2026-10-10

Updated: 2026-10-20

Comments: Accepted at the NeurIPS 2026 Workshop on Foundations of Language Model Security (FLMSec). 10 pages, 4 figures, 3 tables

Code: https://github.com/LassNajmuldeen/balcony-jailbreaking

Project page: https://hamidmousavi0.github.io

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Sources

• AdversariaLLM: A Unified and Modular Toolbox for LLM Robustness Research
• DeepSeek LLM: Scaling Open-Source Language Models with Longtermism
• Flextron: Many-in-One Flexible Large Language Model
• Explaining and Harnessing Adversarial Examples
• The Llama 3 Herd of Models
• MatFormer: Nested Transformer for Elastic Inference
• HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
• Claudini: Autoresearch Discovers State-of-the-Art Adversarial Attack Algorithms for LLMs
• Adversarial Reasoning at Jailbreaking Time
• Fast Adversarial Attacks on Language Models In One GPU Minute
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• TAO-Attack: Toward Advanced Optimization-Based Jailbreak Attacks for Large Language Models
• AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models
• Universal and Transferable Adversarial Attacks on Aligned Language Models

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks