DRE: An Effective Dual-Refined Method for Integrating Small and Large Language Models in Open-Domain Dialogue Evaluation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912414667112448 |
|---|---|
| author | Zhao, Kun Yang, Bohao Tang, Chen Dai, Siyuan Tang, Haoteng Lin, Chenghua Zhan, Liang |
| author_facet | Zhao, Kun Yang, Bohao Tang, Chen Dai, Siyuan Tang, Haoteng Lin, Chenghua Zhan, Liang |
| contents | Large Language Models (LLMs) excel at many tasks but struggle with ambiguous scenarios where multiple valid responses exist, often yielding unreliable results. Conversely, Small Language Models (SLMs) demonstrate robustness in such scenarios but are susceptible to misleading or adversarial inputs. We observed that LLMs handle negative examples effectively, while SLMs excel with positive examples. To leverage their complementary strengths, we introduce SLIDE (Small and Large Integrated for Dialogue Evaluation), a method integrating SLMs and LLMs via adaptive weighting. Building on SLIDE, we further propose a Dual-Refinement Evaluation (DRE) method to enhance SLM-LLM integration: (1) SLM-generated insights guide the LLM to produce initial evaluations; (2) SLM-derived adjustments refine the LLM's scores for improved accuracy. Experiments demonstrate that DRE outperforms existing methods, showing stronger alignment with human judgment across diverse benchmarks. This work illustrates how combining small and large models can yield more reliable evaluation tools, particularly for open-ended tasks such as dialogue evaluation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_04516 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DRE: An Effective Dual-Refined Method for Integrating Small and Large Language Models in Open-Domain Dialogue Evaluation Zhao, Kun Yang, Bohao Tang, Chen Dai, Siyuan Tang, Haoteng Lin, Chenghua Zhan, Liang Computation and Language Large Language Models (LLMs) excel at many tasks but struggle with ambiguous scenarios where multiple valid responses exist, often yielding unreliable results. Conversely, Small Language Models (SLMs) demonstrate robustness in such scenarios but are susceptible to misleading or adversarial inputs. We observed that LLMs handle negative examples effectively, while SLMs excel with positive examples. To leverage their complementary strengths, we introduce SLIDE (Small and Large Integrated for Dialogue Evaluation), a method integrating SLMs and LLMs via adaptive weighting. Building on SLIDE, we further propose a Dual-Refinement Evaluation (DRE) method to enhance SLM-LLM integration: (1) SLM-generated insights guide the LLM to produce initial evaluations; (2) SLM-derived adjustments refine the LLM's scores for improved accuracy. Experiments demonstrate that DRE outperforms existing methods, showing stronger alignment with human judgment across diverse benchmarks. This work illustrates how combining small and large models can yield more reliable evaluation tools, particularly for open-ended tasks such as dialogue evaluation. |
| title | DRE: An Effective Dual-Refined Method for Integrating Small and Large Language Models in Open-Domain Dialogue Evaluation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2506.04516 |