Alignment Flywheel-Guided Suppression Patching for Interactionless Inverse Reinforcement Learning
cs.LG
Submitted: 2026-10-31
Updated: 2026-10-31
Code: https://github.com/decide-ugent/Alignment-Flywheel
Terminology
Sources
- Concrete Problems in AI Safety
- Exploration by Random Network Distillation
- Transformer-Patcher: One Mistake worth One Neuron
- BetaEdit: Null-Space Constrained Sequential Model Editing
- Mixture of Autoencoder Experts Guidance using Unlabeled and Incomplete Data for Exploration in Reinforcement Learning
- The Alignment Flywheel: A Governance-Centric Hybrid MAS for Architecture-Agnostic Safety
- The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models
- Active Reward Machine Inference From Raw State Trajectories
- Partial Identifiability and Misspecification in Inverse Reinforcement 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