AI Alignment via Incentives and Correction
cs.LG, cs.AI
Submitted: 2026-05-02
Updated: 2026-11-12
Terminology
Sources
- Concrete Problems in AI Safety
- Adversarial Reward Auditing for Active Detection and Mitigation of Reward Hacking
- Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision
- Supervising strong learners by amplifying weak experts
- Training Verifiers to Solve Math Word Problems
- Improving Factuality and Reasoning in Language Models through Multiagent Debate
- Reward Tampering Problems and Solutions in Reinforcement Learning: A Causal Influence Diagram Perspective
- Games for AI Control: Models of Safety Evaluations of AI Deployment Protocols
- AI safety via debate
- Scalable agent alignment via reward modeling: a research direction
- Let's Verify Step by Step
- LLM Critics Help Catch LLM Bugs
Related papers
- 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
- (Mis)generalization of Helpful-Only Fine-tuning