Writerslogic improves biomedical text simplification and error detection
Writerslogic at the CLEF 2026 SimpleText Track: Multi-Candidate LLM Simplification and Stacked Complexity Spotting
Computation and Language
Summary
Simplifying complex biomedical texts helps more people understand health information. Writerslogic created a system that makes multiple simpler versions of a sentence and picks the best one using smart rules. They also built a model that spots when information is added wrongly or exaggerated by checking if sentences match trusted sources. Their tools work well on English and other languages using medical reviews. These improvements can help make health documents easier and more reliable.
What this means in practice
- •For healthcare software developers: Create tools that automatically simplify medical documents for patients using multi-candidate language model generation and scoring.
- •For medical ai safety teams: Implement systems that detect incorrect or hallucinated information in biomedical AI outputs by modeling sentence-level grounding to sources.
Authors
David L. Condrey
Abstract
We describe the Writerslogic team's participation in the CLEF 2026 SimpleText shared task, addressing Task 1 (text simplification) and Task 2 (complexity spotting). For Task 1, we develop a multi-candidate generation pipeline using GPT-4o-mini that produces five simplification candidates per sentence at varying temperatures, then selects the best candidate using a reference-free scoring heuristic that rewards compression, source word retention, Cochrane Plain Language Summary vocabulary usage, and lexical simplicity. On Task 1.1 (sentence-level simplification), our Claude Sonnet 4 submission achieves SARI 47.43 and BLEU 14.21, the top-ranked sentence-level system (3rd on the combined Task 1 leaderboard, behind two document-level submissions). For Task 2, we fine-tune a DeBERTa-v3-large NLI model on 350K labeled (source, sentence) pairs, framing hallucination detection as natural language inference. The model reads the most relevant source sentence as premise and the candidate as hypothesis, directly learning to distinguish grounded from hallucinated content. On Task 2.1 (binary overgeneration identification), our fine-tuned DeBERTa system achieves 0.8081 document-level macro F1 (0.8085 in our best ensemble), the top-ranked entry within the identification track and 2nd among teams overall, behind AIIR Lab (0.8197). On Task 2.2 (multi-class error classification), our best submission reaches 0.804 multiclass accuracy, ranking 2nd among unique teams behind AIIR Lab (0.827). We evaluate both tasks on English and multilingual biomedical text from Cochrane systematic reviews.