Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2604.09629 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913169338793984 |
|---|---|
| author | Ajayi, Edward Mitra, Prasenjit |
| author_facet | Ajayi, Edward Mitra, Prasenjit |
| contents | Humor generation poses a significant challenge for Large Language Models (LLMs), because their standard training objective (next-token prediction) inherently conflicts with the surprise and incongruity required for comedy. To bridge this gap, we introduce the Cognitive Synergy Framework, a methodology for generating highquality humor data inspired by psychological theories of humor. Utilizing a Mixtureof-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 produces a theory-grounded dataset, which we use to fine-tune a 7B-parameter student model. We further evaluate two alignment strategies, Direct Preference Optimization (DPO) and an offline group-relative variant O-GRPO, finding that neither improves over SFT. However, our 7B HumorGen model variants significantly outperform larger instruction-tuned baselines and achieve top-tier open-weight performance while remaining competitive with frontier proprietary systems. These results suggest that cognitively driven data curation is more critical than alignment algorithms or model scale for humor generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_09629 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | HumorGen: Cognitive Synergy for Humor Generation in Large Language Models via Persona-Based Distillation Ajayi, Edward Mitra, Prasenjit Computation and Language Humor generation poses a significant challenge for Large Language Models (LLMs), because their standard training objective (next-token prediction) inherently conflicts with the surprise and incongruity required for comedy. To bridge this gap, we introduce the Cognitive Synergy Framework, a methodology for generating highquality humor data inspired by psychological theories of humor. Utilizing a Mixtureof-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 produces a theory-grounded dataset, which we use to fine-tune a 7B-parameter student model. We further evaluate two alignment strategies, Direct Preference Optimization (DPO) and an offline group-relative variant O-GRPO, finding that neither improves over SFT. However, our 7B HumorGen model variants significantly outperform larger instruction-tuned baselines and achieve top-tier open-weight performance while remaining competitive with frontier proprietary systems. These results suggest that cognitively driven data curation is more critical than alignment algorithms or model scale for humor generation. |
| title | HumorGen: Cognitive Synergy for Humor Generation in Large Language Models via Persona-Based Distillation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2604.09629 |