4Hammer: a board-game reinforcement learning environment for the hour long time frame

Massimo Fioravanti, Giovanni Agosta

Large Language Models (LLMs) have demonstrated strong performance on tasks with short time frames, but struggle with tasks requiring longer durations. While datasets covering exten…