GenCeption: Evaluate Multimodal LLMs with Unlabeled Unimodal Data

Code

• github.com/llcresearch/genception
  OfficialIn papernone★ 1

Abstract

Multimodal Large Language Models (MLLMs) are typically assessed using expensive annotated multimodal benchmarks, which often lag behind the rapidly evolving demands of MLLM evaluation. This paper outlines and validates GenCeption, a novel, annotation-free evaluation method that requires only unimodal data to measure inter-modality semantic coherence and inversely assesses MLLMs' tendency to hallucinate. This approach eliminates the need for costly data annotation, minimizes the risk of training data contamination, results in slower benchmark saturation, and avoids the illusion of emerging abilities. Inspired by the DrawCeption game, GenCeption begins with a non-textual sample and proceeds through iterative description and generation steps. The semantic drift across iterations is quantified using the GC@T metric. Based on the GenCeption method, we establish the MMECeption benchmark for evaluating Vision LLMs (VLLMs), and compare performance of several popular VLLMs and human annotators. Our empirical results validate GenCeption's effectiveness, demonstrating strong correlations with established VLLM benchmarks. VLLMs still significantly lack behind human performance and struggle especially with text-intensive tasks.

Tasks

Reproductions