Self-playing Adversarial Language Game Enhances LLM Reasoning

arXiv:2404.10642v1 Announce Type: cross
Abstract: We explore the self-play training procedure of large language models (LLMs) in a two-player adversarial language game called Adversarial Taboo. In this game, an attacker and a defender communicate with respect to a target word only visible to the attacker. The attacker aims to induce the defender to utter the target word unconsciously, while the defender tries to infer the target word from the attacker's utterances. To win the game, both players should have sufficient …

adversarial arxiv cs.cl cs.lg game language llm llm reasoning playing reasoning type