Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2507.15752 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911067811086336 |
|---|---|
| author | Zhu, Ruizhe Zhu, Hao Li, Yaxuan Zhou, Syang Cai, Shijing Lazuka, Malgorzata Ash, Elliott |
| author_facet | Zhu, Ruizhe Zhu, Hao Li, Yaxuan Zhou, Syang Cai, Shijing Lazuka, Malgorzata Ash, Elliott |
| contents | Collecting human-chatbot dialogues typically demands substantial manual effort and is time-consuming, which limits and poses challenges for research on conversational AI. In this work, we propose DialogueForge - a framework for generating AI-simulated conversations in human-chatbot style. To initialize each generated conversation, DialogueForge uses seed prompts extracted from real human-chatbot interactions. We test a variety of LLMs to simulate the human chatbot user, ranging from state-of-the-art proprietary models to small-scale open-source LLMs, and generate multi-turn dialogues tailored to specific tasks. In addition, we explore fine-tuning techniques to enhance the ability of smaller models to produce indistinguishable human-like dialogues. We evaluate the quality of the simulated conversations and compare different models using the UniEval and GTEval evaluation protocols. Our experiments show that large proprietary models (e.g., GPT-4o) generally outperform others in generating more realistic dialogues, while smaller open-source models (e.g., Llama, Mistral) offer promising performance with greater customization. We demonstrate that the performance of smaller models can be significantly improved by employing supervised fine-tuning techniques. Nevertheless, maintaining coherent and natural long-form human-like dialogues remains a common challenge across all models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_15752 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DialogueForge: LLM Simulation of Human-Chatbot Dialogue Zhu, Ruizhe Zhu, Hao Li, Yaxuan Zhou, Syang Cai, Shijing Lazuka, Malgorzata Ash, Elliott Computation and Language Artificial Intelligence Collecting human-chatbot dialogues typically demands substantial manual effort and is time-consuming, which limits and poses challenges for research on conversational AI. In this work, we propose DialogueForge - a framework for generating AI-simulated conversations in human-chatbot style. To initialize each generated conversation, DialogueForge uses seed prompts extracted from real human-chatbot interactions. We test a variety of LLMs to simulate the human chatbot user, ranging from state-of-the-art proprietary models to small-scale open-source LLMs, and generate multi-turn dialogues tailored to specific tasks. In addition, we explore fine-tuning techniques to enhance the ability of smaller models to produce indistinguishable human-like dialogues. We evaluate the quality of the simulated conversations and compare different models using the UniEval and GTEval evaluation protocols. Our experiments show that large proprietary models (e.g., GPT-4o) generally outperform others in generating more realistic dialogues, while smaller open-source models (e.g., Llama, Mistral) offer promising performance with greater customization. We demonstrate that the performance of smaller models can be significantly improved by employing supervised fine-tuning techniques. Nevertheless, maintaining coherent and natural long-form human-like dialogues remains a common challenge across all models. |
| title | DialogueForge: LLM Simulation of Human-Chatbot Dialogue |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2507.15752 |