Judge Before Answer: Can MLLM Discern the False Premise in Question?

2025-10-12

AI Summary

Current multimodal large language models (MLLMs) exhibit poor performance in detecting false premises within image-text pairs, while existing benchmarks lack fine-grained categorization and comprehensive coverage. Method: We introduce the first systematic benchmark for false premise identification, hierarchically classifying false premises into three broad categories and thirteen fine-grained subtypes, and propose an automated, multimodal data construction pipeline. Furthermore, we design a dedicated enhancement framework integrating prompt learning and supervised fine-tuning to improve model robustness. Contribution/Results: Extensive experiments reveal that state-of-the-art MLLMs achieve low average accuracy (<40%) on this benchmark, confirming its difficulty and diagnostic value. Our enhancement framework yields substantial improvements, boosting accuracy by 12.7–28.3 percentage points. This work establishes a new paradigm and practical toolkit for evaluating and improving the logical reliability of MLLM reasoning.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Computer Vision: Multi-modal Vision

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Multimodal large language models (MLLMs) have witnessed astonishing advancements in recent years. Despite these successes, MLLMs remain vulnerable to flase premise problems. However, existing benchmarks targeting this issue are limited in scope: they often lack fine-grained categorization, exhibit insufficient coverage, and thus fail to provide a rigorous evaluation of the ability of models to recognize false premises. To bridge this gap, we introduce a fully automated pipeline for constructing a comprehensive benchmark of false premise questions. Our method systematically categorizes the premises into three main types and thirteen subtypes according to the abilities required to identify the premises, resulting in the JBA dataset.Results show current MLLMs still struggle with false premise recognition. Building upon this benchmark, we further propose a recognition enhancement framework tailored to strengthen the robustness of MLLMs to detect false premises. Extensive experiments demonstrate that models trained with our framework achieve significant improvements in false premise recognition.

Problem

• MLLMs struggle to identify false premises in questions
• Existing benchmarks lack comprehensive categorization and coverage
• Developing automated pipeline to enhance false premise recognition

Innovation

• Automated pipeline constructs comprehensive false premise benchmark
• Systematically categorizes premises into three main types
• Recognition enhancement framework strengthens MLLM false premise detection