Post-training LLMs

Start with a pretrained model that only predicts text. End understanding every lever that turns it into an aligned, reasoning assistant — RLHF, DPO, the GRPO family, and the test-time-compute frontier.

• Training language models to follow instructions with human feedback
• Constitutional AI: Harmlessness from AI Feedback
• Direct Preference Optimization: Your Language Model is Secretly a Reward Model
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
• DAPO: An Open-Source LLM Reinforcement Learning System at Scale
• Group Sequence Policy Optimization
• DeepSeek-R1
• Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
• s1: Simple test-time scaling
• Tülu 3: Pushing Frontiers in Open Language Model Post-Training