Reinforcement Learning for Design Taste — Jasper Lu
jasperlu.com · saved by 1 readers
Training Qwen3.8-27B on MiMo's webdev RL environments, and how the choice of LLM judge shapes a model's design taste.
Training Qwen3.8-27B on MiMo's webdev RL environments, and how the choice of LLM judge shapes a model's design taste.