Calibrating LLMs with Preference Optimization on Thought Trees for Generating Rationale in Science Question Scoring

Jiazheng Li, Hainiu Xu, Zhaoyue Sun, Yuxiang Zhou, David West, Cesare Aloisi, Yulan He


Abstract
Generating rationales that justify scoring decisions has been a promising way to facilitate explainability in automated scoring systems. However, existing methods do not match the accuracy of classifier-based methods. Plus, the generated rationales often contain hallucinated information. To address these issues, we propose a novel framework capable of generating more faithful rationales and, more importantly, matching performance with classifier-based black-box scoring systems. We first mimic the human assessment process by querying Large Language Models (LLMs) to generate a thought tree. We then summarise intermediate assessment decisions from each thought tree path for creating synthetic rationale data and rationale preference data. Finally, we utilise the generated synthetic data to calibrate LLMs through a two-step training process: supervised fine-tuning and preference optimization. Extensive experimental results demonstrate that our framework achieves a 38% assessment performance improvement in the QWK score compared to prior work while producing higher-quality rationales, as recognised by human evaluators and LLMs. Our work sheds light on the effectiveness of performing preference optimization using synthetic preference data obtained from thought tree paths. Data and code are available at: https://github.com/lijiazheng99/thought_tree_assessment.
Anthology ID:
2024.findings-emnlp.313
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2024
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5452–5479
Language:
URL:
https://aclanthology.org/2024.findings-emnlp.313
DOI:
10.18653/v1/2024.findings-emnlp.313
Bibkey:
Cite (ACL):
Jiazheng Li, Hainiu Xu, Zhaoyue Sun, Yuxiang Zhou, David West, Cesare Aloisi, and Yulan He. 2024. Calibrating LLMs with Preference Optimization on Thought Trees for Generating Rationale in Science Question Scoring. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 5452–5479, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Calibrating LLMs with Preference Optimization on Thought Trees for Generating Rationale in Science Question Scoring (Li et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-emnlp.313.pdf