MetaRuleGPT: Recursive Numerical Reasoning of Language Models Trained with Simple Rules

ReLM
LRM
Re-assign community