Hallucination Detection and Mitigation in Scientific Text Simplification using Ensemble Approaches: DS@GT at CLEF 2025 SimpleText

TL;DR

This paper presents an ensemble approach combining multiple models to detect and mitigate hallucinations and distortions in scientific text simplification, improving the reliability of generated simplified texts.

Contribution

It introduces a novel ensemble framework integrating BERT, semantic similarity, NLI, and LLM reasoning for hallucination detection in scientific text simplification.

Findings

Enhanced detection accuracy through ensemble methods

Effective post-editing with LLM-based revisions

Improved robustness against information distortion

Abstract

In this paper, we describe our methodology for the CLEF 2025 SimpleText Task 2, which focuses on detecting and evaluating creative generation and information distortion in scientific text simplification. Our solution integrates multiple strategies: we construct an ensemble framework that leverages BERT-based classifier, semantic similarity measure, natural language inference model, and large language model (LLM) reasoning. These diverse signals are combined using meta-classifiers to enhance the robustness of spurious and distortion detection. Additionally, for grounded generation, we employ an LLM-based post-editing system that revises simplifications based on the original input texts.