A Single Suffix to Break Them All: Basin-Aware Jailbreaks for Merged Model Families
cs.LG, cs.CL
Submitted: 2026-08-27
Updated: 2026-08-27
Code: https://github.com/sahil280114/codealpaca
Terminology
Sources
- Qwen Technical Report
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- DeepSeek LLM: Scaling Open-Source Language Models with Longtermism
- Language Models are Few-Shot Learners
- Jailbreaking Black Box Large Language Models in Twenty Queries
- Training Verifiers to Solve Math Word Problems
- DELLA-Merging: Reducing Interference in Model Merging through Magnitude-Based Sampling
- BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
- BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors
- Baseline Defenses for Adversarial Attacks Against Aligned Language Models
- Mistral 7B
- AmpleGCG: Learning a Universal and Transferable Generative Model of Adversarial Suffixes for Jailbreaking Both Open and Closed LLMs
- "Do Anything Now": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Models
- One Leak Away: How Pretrained Model Exposure Amplifies Jailbreak Risks in Finetuned LLMs
- Gemma: Open Models Based on Gemini Research and Technology
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- Mitigating Jailbreaks with Intent-Aware LLMs
- Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
- On the Emergence of Cross-Task Linearity in the Pretraining-Finetuning Paradigm
- 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: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks