Selective Safety Steering via Value-Filtered Decoding
cs.LG
Submitted: 2026-05-14
Updated: 2026-11-02
Terminology
Sources
- Fine-Tuning Language Models from Human Preferences
- Proximal Policy Optimization Algorithms
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- SafeDPO: A Simple Approach to Direct Preference Optimization with Enhanced Safety
- Creativity Has Left the Chat: The Price of Debiasing Language Models
- Controlled Decoding from Language Models
- Value Augmented Sampling for Language Model Alignment and Personalization
- STARS: Synchronous Token Alignment for Robust Supervision in Large Language Models
- Conformal Policy Control
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- Constitutional AI: Harmlessness from AI Feedback
- A General Language Assistant as a Laboratory for Alignment
- Tradeoffs Between Alignment and Helpfulness in Language Models with Steering Methods
- E-valuator: Reliable Agent Verifiers with Sequential Hypothesis Testing
- Mistral 7B
- The Llama 3 Herd of Models
- A Primer on PAC-Bayesian Learning
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis