This repository contains the experimental code for the ReAct (Reasoning + Acting) prompting paper published at ICLR 2023. The work investigates how language models like GPT-3 and PaLM can synergize internal reasoning with external actions to improve task performance. The notebooks run experiments on HotpotQA, FEVER, AlfWorld, and WebShop datasets, measuring accuracy and success rates across reasoning and decision-making tasks.