HumorGen: Cognitive Synergy for Humor Generation in Large Language Models via Persona-Based Distillation

2026-03-19

AI Summary

This work addresses the challenge that large language models struggle to generate high-quality humor due to a fundamental misalignment between their training objectives and the incongruity and surprise essential to humor. To overcome this, the authors propose a cognitive synergy framework that integrates psychological theories of humor into data construction for the first time. They synthesize diverse humorous content using a Mixture-of-Thought strategy driven by six cognitive personas—such as the Absurdist and the Cynic—and employ persona-guided data distillation followed by supervised fine-tuning to train HumorGen, a 7B-parameter model. Experiments demonstrate that this approach significantly outperforms larger instruction-tuned baselines, achieves state-of-the-art performance among open-source models, and rivals leading closed-source systems, thereby validating the critical role of cognitive synergy in data distillation for effective humor generation.

Abstract

Humor generation poses a significant challenge for Large Language Models (LLMs), because their standard training objective - predicting the most likely next word - inherently conflicts with the surprise and incongruity needed for comedy. To bridge this gap, we introduce the Cognitive Synergy Framework, a theoretically grounded methodology for generating high-quality humor data inspired by psychological theories of humor. Utilizing a Mixture-of-Thought (MoT) approach, we deploy six cognitive personas (e.g., The Absurdist, The Cynic) to synthesize diverse comedic perspectives for a given prompt. This framework creates a theoretically grounded dataset, which we use to fine-tune a 7B-parameter student model. We compare Direct Preference Optimization (DPO) and a novel Offline Group Relative Policy Optimization (O-GRPO); our 7B model significantly outperforms larger instruction-tuned baselines and achieves performance competitive with state-of-the-art proprietary models. We find that cognitive-driven data curation is far more critical than alignment algorithms or model scale for humor generation. Code and data will be available upon publication.

Problem

• humor generation
• large language models
• cognitive synergy
• incongruity
• surprise

Innovation

• Cognitive Synergy Framework
• Mixture-of-Thought
• Persona-Based Distillation
• Humor Generation
• Theory-Grounded Dataset