Algorithm of Thoughts: Enhancing Exploration of Ideas in Large Language ModelsCurrent literature, aiming to surpass the "Chain-of-Thought" approach, oftenresorts to an external modus operandi involving halting, modifying, and thenresuming the generation process to boost Large Language Models' (LLMs)reasoning capacities. This mode escalates the number of query requests, leadingto increased costs, memory, and computational overheads. Addressing this, wepropose the Algorithm of Thoughts -- a novel strategy that propels LLMs throughalgorithmic reasoning pathways, pioneering a new mode of in-context learning.By employing algorithmic examples, we exploit the innate recurrence dynamics ofLLMs, expanding their idea exploration with merely one or a few queries. Ourtechnique outperforms earlier single-query methods and stands on par with arecent multi-query strategy that employs an extensive tree search algorithm.Intriguingly, our results suggest that instructing an LLM using an algorithmcan lead to performance surpassing that of the algorithm itself, hinting atLLM's inherent ability to weave its intuition into optimized searches. We probeinto the underpinnings of our method's efficacy and its nuances in application.arxiv.org