Palette: A Modular, Controllable, and Efficient Framework for On-demand Authorized Safety Alignment Relaxation in LLMs

arXiv:2605.24154 · cs.AI, cs.SE · Submitted 2026-05-22 · Read on arXiv

cs.AI, cs.SE

Submitted: 2026-05-22

Updated: 2026-10-05

Terminology

• Alignment — 2× in this paper · explained in 1 episode
• Foundation Model — 2× in this paper · explained in 3 episodes
• Foundation Models — 2× in this paper · explained in 7 episodes
• Adaptation — 1× in this paper · explained in 1 episode
• Benchmark — 1× in this paper · explained in 2 episodes
• Inference-Time Steering — 1× in this paper · explained in 2 episodes
• Steering — 1× in this paper · explained in 2 episodes

Sources

• MITS: Enhanced Tree Search Reasoning for LLMs via Pointwise Mutual Information
• Harmony in Divergence: Towards Fast, Accurate, and Memory-efficient Zeroth-order LLM Fine-tuning
• Mitigating Hallucination Through Theory-Consistent Symmetric Multimodal Preference Optimization
• Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!
• Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models
• PEFT-as-an-Attack! Jailbreaking Language Models during Federated Parameter-Efficient Fine-Tuning
• Emerging Safety Attack and Defense in Federated Instruction Tuning of Large Language Models
• Q-realign: Piggybacking Realignment on Quantization for Safe and Efficient LLM Deployment
• DAPO: An Open-Source LLM Reinforcement Learning System at Scale
• Controllable Safety Alignment: Inference-Time Adaptation to Diverse Safety Requirements
• Programming Refusal with Conditional Activation Steering
• The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
• Steering Llama 2 via Contrastive Activation Addition
• HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
• Universal and Transferable Adversarial Attacks on Aligned Language Models
• BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset
• The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• The Llama 3 Herd of Models
• Qwen3 Technical Report

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems