Mid-training 1

• Post-training degli LLM: fine-tuning e reinforcement learning Sep 30, 2026