Skip to content
All tags

#grpo

2 posts
ai deep-dive 認識 AI 模型

Three RL Post-Training Playbooks: How Ornith, Nous Research, and MiniMax Built Dark Horse Models

Three non-big-lab teams used different RL post-training strategies to produce benchmark dark horses in 2026: Ornith's self-improvement loop (GRPO), Nous Research's DataForge + Atropos execution-reward RL, and MiniMax's massive-scale RL across 200K real environments. Different strengths, but one shared proof point: post-training RL matters more than pretraining scale.

#reinforcement-learning#grpo#post-training#self-improvement#open-source#agentic-coding#fine-tuning
ai deep-dive Reading Stanford CS336

CS336 Lecture 16: RLVR Scales Reasoning with Verifiable Rewards, but GRPO Is Not Free PPO

Lecture 16 moves from PPO to GRPO and RLVR. Math, code, and environment outcomes provide scalable rewards and avoid some preference-model overoptimization, but group-normalized advantages introduce difficulty and length bias while rollout infrastructure becomes the dominant cost.

#cs336#rlvr#grpo#reasoning#reinforcement-learning