Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2506.15131 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909979454210048 |
|---|---|
| author | Lee, Jing Yang Lee, Kong-Aik Gan, Woon-Seng |
| author_facet | Lee, Jing Yang Lee, Kong-Aik Gan, Woon-Seng |
| contents | Open-domain Dialogue (OD) exhibits a one-to-many (o2m) property, whereby multiple appropriate responses exist for a single dialogue context. Despite prior research showing that modeling this property boosts response diversity, most modern LLM-based dialogue agents do not explicitly do so. In this work, we model the o2m property of OD in LLMs by decomposing OD generation into two key tasks: Multi-Response Generation (MRG) and Preference-based Selection (PS), which entail generating a set of n semantically and lexically diverse high-quality responses for a given dialogue context, followed by selecting a single response based on human preference, respectively. To facilitate MRG and PS, we introduce o2mDial, a dialogue corpus explicitly designed to capture the o2m property by featuring multiple plausible responses for each context. Leveraging o2mDial, we propose new in-context learning and instruction-tuning strategies, as well as novel evaluation metrics for MRG, alongside a model-based approach for PS. Empirical results demonstrate that applying the proposed two-stage framework to smaller LLMs for OD generation enhances overall response diversity while maintaining contextual coherence, improving response quality by up to 90%, bringing them closer to the performance of larger models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_15131 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Modeling the One-to-Many Property in Open-Domain Dialogue with LLMs Lee, Jing Yang Lee, Kong-Aik Gan, Woon-Seng Computation and Language Artificial Intelligence Open-domain Dialogue (OD) exhibits a one-to-many (o2m) property, whereby multiple appropriate responses exist for a single dialogue context. Despite prior research showing that modeling this property boosts response diversity, most modern LLM-based dialogue agents do not explicitly do so. In this work, we model the o2m property of OD in LLMs by decomposing OD generation into two key tasks: Multi-Response Generation (MRG) and Preference-based Selection (PS), which entail generating a set of n semantically and lexically diverse high-quality responses for a given dialogue context, followed by selecting a single response based on human preference, respectively. To facilitate MRG and PS, we introduce o2mDial, a dialogue corpus explicitly designed to capture the o2m property by featuring multiple plausible responses for each context. Leveraging o2mDial, we propose new in-context learning and instruction-tuning strategies, as well as novel evaluation metrics for MRG, alongside a model-based approach for PS. Empirical results demonstrate that applying the proposed two-stage framework to smaller LLMs for OD generation enhances overall response diversity while maintaining contextual coherence, improving response quality by up to 90%, bringing them closer to the performance of larger models. |
| title | Modeling the One-to-Many Property in Open-Domain Dialogue with LLMs |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2506.15131 |