OUNLP at TSAR 2025 Shared Task: Multi-Round Text Simplifier via Code Generation
This is an incremental improvement for text simplification systems, addressing the specific problem of readability control for users needing simplified text.
The paper tackled readability-controlled text simplification by proposing multi-round methods using GPT-4o, based on the finding that performance relates to the gap between source and target CEFR levels, with their system ranking 7th out of 20 teams in the TSAR-2025 Shared Task.
This paper describes the OUNLP system submitted to the TSAR-2025 Shared Task (Alva-Manchego et al., 2025), designed for readability-controlled text simplification using LLM-prompting-based generation. Based on the analysis of prompt-based text simplification methods, we discovered an interesting finding that text simplification performance is highly related to the gap between the source CEFR (Arase et al., 2022) level and the target CEFR level. Inspired by this finding, we propose two multi-round simplification methods and generate them via GPT-4o: rule-based simplification (MRS-Rule) and jointly rule-based LLM simplification (MRS-Joint). Our submitted systems ranked 7 out of 20 teams. Later improvements with MRS-Joint show that taking the LLM simplified candidates as the starting point could further boost the multi-round simplification performance.