Primitives and Paths: Predicting What Post-Training Can Do from What the Base Model Already Has
cs.LG, cs.AI
Submitted: 2026-11-08
Updated: 2026-11-08
Terminology
Sources
- The Coverage Principle: How Pre-Training Enables Post-Training
- SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training
- Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs
- DeReason: A Difficulty-Aware Curriculum Improves Decoupled SFT-then-RL Training for General Reasoning
- Quagmires in SFT-RL Post-Training: When High SFT Scores Mislead and What to Use Instead
- Learning What Reinforcement Learning Can't: Interleaved Online Fine-Tuning for Hardest Questions
- Decomposing Elements of Problem Solving: What "Math" Does RL Teach?
- From $f(x)$ and $g(x)$ to $f(g(x))$: LLMs Learn New Skills in RL by Composing Old Ones
- Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
- On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
- Echo Chamber: RL Post-training Amplifies Behaviors Learned in Pretraining
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