Innovative Thinking, Infinite Humor: Humor Research of Large Language Models through Structured Thought Leaps

TL;DR

This paper introduces LoL, a framework that enhances humor generation in large language models by integrating external knowledge and multi-hop reasoning, thereby improving their creative and judgment capabilities.

Contribution

The paper proposes a novel LoL framework that combines instruction-evolution and reinforcement learning to improve humor reasoning and generation in large language models.

Findings

Enhanced humor reasoning and generation in LLMs.

Improved judgment ability and creative capacity of models.

Deeper understanding of LLMs' creative capabilities.

Abstract

Humor is previously regarded as a gift exclusive to humans for the following reasons. Humor is a culturally nuanced aspect of human language, presenting challenges for its understanding and generation. Humor generation necessitates a multi-hop reasoning process, with each hop founded on proper rationales. Although many studies, such as those related to GPT-o1, focus on logical reasoning with reflection and correction, they still fall short in humor generation. Due to the sparsity of the knowledge graph in creative thinking, it is arduous to achieve multi-hop reasoning. Consequently, in this paper, we propose a more robust framework for addressing the humor reasoning task, named LoL. LoL aims to inject external information to mitigate the sparsity of the knowledge graph, thereby enabling multi-hop reasoning. In the first stage of LoL, we put forward an automatic instruction-evolution method to incorporate the deeper and broader thinking processes underlying humor. Judgment-oriented instructions are devised to enhance the model's judgment capability, dynamically supplementing and updating the sparse knowledge graph. Subsequently, through reinforcement learning, the reasoning logic for each online-generated response is extracted using GPT-4o. In this process, external knowledge is re-introduced to aid the model in logical reasoning and the learning of human preferences. Finally, experimental results indicate that the combination of these two processes can enhance both the model's judgment ability and its generative capacity. These findings deepen our comprehension of the creative capabilities of large language models (LLMs) and offer approaches to boost LLMs' creative abilities for cross-domain innovative applications.