1
SUPERNOVA: Eliciting General Reasoning in LLMs with Reinforcement Learning on Natural Instructions
用强化学习在自然指令上训练，让大模型学会通用推理，SUPERNOVA方法开启新思路。
arXiv:2604.08477v2 Announce Type: replace Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has substantially improved reasoning in form…