CardiffNLP at CLEARS-2025: Prompting Large Language Models for Plain Language and Easy-to-Read Text Rewriting

2025-08-05

AI Summary

This study addresses Spanish text simplification, targeting two complementary rewriting objectives—plain language adaptation and readability enhancement—as part of the CLEARS-2025 shared task. We propose a fine-tuning-free, lightweight prompt engineering framework that integrates few-shot learning, diverse prompt templates, and in-model knowledge guidance. The approach is systematically evaluated on Gemma-3 and LLaMA-3.2 to assess how prompt design influences simplification quality. Our method drastically reduces computational overhead while achieving third place in Subtask 1 (plain language rewriting) and second place in Subtask 2 (readability optimization). These results validate the effectiveness and generalizability of structured prompt engineering for Spanish readability transformation. Moreover, the framework establishes a novel paradigm for resource-efficient text simplification in low-resource languages, demonstrating strong performance without parameter updates or domain-specific fine-tuning.

Technology Category

• Natural Language Processing: Prompt Engineering / Prompting
• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text

Abstract

This paper details the CardiffNLP team's contribution to the CLEARS shared task on Spanish text adaptation, hosted by IberLEF 2025. The shared task contained two subtasks and the team submitted to both. Our team took an LLM-prompting approach with different prompt variations. While we initially experimented with LLaMA-3.2, we adopted Gemma-3 for our final submission, and landed third place in Subtask 1 and second place in Subtask 2. We detail our numerous prompt variations, examples, and experimental results.

Problem

• Adapting Spanish text for plain language using LLMs
• Rewriting text for easy-to-read comprehension
• Exploring prompt variations for LLM-based text simplification

Innovation

• Used LLM-prompting approach for text adaptation
• Experimented with LLaMA-3.2 and Gemma-3
• Employed numerous prompt variations and examples