TruthRL: Incentivizing Truthful LLMs via Reinforcement Learning - Explained Simply | ArXiv Explained