Hal-Eval: A Universal and Fine-grained Hallucination Evaluation Framework for Large Vision Language Models
This addresses the issue of evaluating complex hallucinations in LVLMs for researchers and developers, though it is incremental as it builds on existing taxonomies.
The paper tackles the problem of hallucinations in Large Vision Language Models by introducing a new category called Event Hallucination and creating a fine-grained evaluation framework. The result is a benchmark that assesses LVLMs' ability to handle a broad spectrum of hallucinations, with code and data to be released.
Large Vision Language Models exhibit remarkable capabilities but struggle with hallucinations inconsistencies between images and their descriptions. Previous hallucination evaluation studies on LVLMs have identified hallucinations in terms of objects, attributes, and relations but overlooked complex hallucinations that create an entire narrative around a fictional entity. In this paper, we introduce a refined taxonomy of hallucinations, featuring a new category: Event Hallucination. We then utilize advanced LLMs to generate and filter fine grained hallucinatory data consisting of various types of hallucinations, with a particular focus on event hallucinations, laying the groundwork for integrating discriminative and generative evaluation methods within our universal evaluation framework. The proposed benchmark distinctively assesses LVLMs ability to tackle a broad spectrum of hallucinations, making it a reliable and comprehensive tool for gauging LVLMs efficacy in handling hallucinations. We will release our code and data.