RL-ATTACK: A Generalized Black-Box Adversarial Attack on Language Models via Reinforcement Learning
cs.CL, cs.CR
Submitted: 2026-11-05
Updated: 2026-11-05
Comments: Accepted as a regular paper at IEEE EdgeCom 2026
Code: https://github.com/taikunudel/RL-ATTACK
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- BERT: A Review of Applications in Natural Language Processing and Understanding
- RoBERTa: A Robustly Optimized BERT Pretraining Approach
- The Llama 3 Herd of Models
- Gemini: A Family of Highly Capable Multimodal Models
- GPT-4 Technical Report
- TextBugger: Generating Adversarial Text Against Real-world Applications
- Deep Reinforcement Learning For Sequence to Sequence Models
- A Deep Reinforced Model for Abstractive Summarization
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data