$T^2$ of Thoughts: Temperature Tree Elicits Reasoning in Large Language Models
This work addresses the adaptability of LLMs in dynamic environments, offering an incremental enhancement to reasoning capabilities for AI applications.
The paper tackled the problem of static reasoning in large language models by introducing Temperature Tree ($T^2$) prompting with a heuristic algorithm, $T^2$ of Thoughts ($T^2oT$), which dynamically adjusts search parameters like temperature to enhance decision-making without extra computational cost, resulting in improvements in accuracy, multi-solution generation, and text generation quality.
Large Language Models (LLMs) have emerged as powerful tools in artificial intelligence, especially in complex decision-making scenarios, but their static problem-solving strategies often limit their adaptability to dynamic environments. We explore the enhancement of reasoning capabilities in LLMs through Temperature Tree ($T^2$) prompting via a heuristic algorithm, termed as $T^2$ of Thoughts ($T^2oT$). The primary focus is on enhancing decision-making processes by dynamically adjusting search parameters, especially temperature, to improve accuracy without increasing computational demands. We empirically validate that our hybrid $T^2oT$ approach yields enhancements in, single-solution accuracy, multi-solution generation and text generation quality. Our findings suggest that while dynamic search depth adjustments based on temperature can yield mixed results, a fixed search depth, when coupled with adaptive capabilities of $T^2oT$, provides a more reliable and versatile problem-solving strategy. This work highlights the potential for future explorations in optimizing algorithmic interactions with foundational language models, particularly illustrated by our development for the Game of 24 and Creative Writing tasks.