paper / russell-stuart / May 9 / failedRobust and Diverse Multi-Agent Learning via Rational Policy Gradient
paper / russell-stuart / May 9 / failedSynthetic Error Injection Fails to Elicit Self-Correction In Language Models