LLM Alignment--Utility Asymmetry under Semantic-Preserving Transformations
cs.CL, cs.AI
Submitted: 2026-09-26
Updated: 2026-09-26
Terminology
Sources
- Aligning Large Language Models with Human: A Survey
- Instruction Tuning for Large Language Models: A Survey
- Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback
- Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
- No, of Course I Can! Deeper Fine-Tuning Attacks That Bypass Token-Level Safety Mechanisms
- Attack via Overfitting: 10-shot Benign Fine-tuning to Jailbreak LLMs
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs
- LARGO: Latent Adversarial Reflection through Gradient Optimization for Jailbreaking LLMs
- AdvPrefix: An Objective for Nuanced LLM Jailbreaks
- FlipAttack: Jailbreak LLMs via Flipping
- Best-of-N Jailbreaking
- Endless Jailbreaks with Bijection Learning
- Hiding in Plain Sight: A Steganographic Approach to Stealthy LLM Jailbreaks
- The Llama 3 Herd of Models
- Gemma: Open Models Based on Gemini Research and Technology
- Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement
- Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge
- Measuring Massive Multitask Language Understanding
- Scaling and evaluating sparse autoencoders
- The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering