Synthetic Data Generation for Training Diversified Commonsense Reasoning Models

arXiv:2603.18361v2 Announce Type: replace Abstract: Conversational agents are required to respond to their users not only with high quality (i.e. commonsense bearing) responses, but also considering multiple plausible alternative scenarios, reflecting the diversity in their responses. Despite the growing need to train diverse commonsense generators, the progress of this line of work has been significantly hindered by the lack of large-scale high-quality diverse commonsense training datasets. Due to the high annotation costs, existing Generative Commonsense Reasoning (GCR) datasets are created using a small number of human annotators, covering only a narrow set of commonsense scenarios. To address this training resource gap, we propose a two-stage method to create the first-ever synthetic dataset CommonSyn for diversified (GCR). The model fine-tuned on our synthetic data jointly increase both generation diversity and quality compared with vanilla models and the model fine-tuned on human-crafted dataset across different size Large Language Models (LLMs)

Related

• How Can We Synthesize High-Quality Pretraining Data? A Systematic Study of Prompt Design, Generator Model, and Source Data
• KCS: Diversify Multi-hop Question Generation with Knowledge Composition Sampling
• Think Multilingual, Not Harder: A Data-Efficient Framework for Teaching Reasoning Models to Code-Switch
• Physical Commonsense Reasoning for Lower-Resourced Languages and Dialects: a Study on Basque

Source: arXiv cs.CL | 2026-04-21