There are competing views on whether RL can genuinely improve base model's performance (e.g., pass@128). The answer is both yes and no, largely depending on the interplay between pre-training, mid-training, and RL. We trained a few hundreds of GPT-2 scale LMs on synthetic GSM-like reasoning data from scratch. Here are what we found: 🧵
RL Impact on Base Model Performance: Pre-training and Mid-training Interplay
By
–
