POPE
Polling-based Object Probing Evaluation (POPE) is a benchmark for evaluating object hallucination in Large Vision-Language Models (LVLMs). POPE addresses the problem where LVLMs generate objects inconsistent with target images by using a polling-based query method that asks yes/no questions about object presence in images, providing more stable and flexible evaluation of object hallucination.
Phi-3.5-vision-instruct from Microsoft currently leads the POPE leaderboard with a score of 0.861 across 2 evaluated AI models.
What POPE measures
POPE is a multimodal benchmark that evaluates large language models on multimodal, safety, and vision tasks. LLM Stats tracks 2 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.9, with the leader reaching 0.9.
Compare leaders on the best AI for multimodal, best AI for safety and best AI for vision leaderboards.
Publication
- Paper
- Evaluating Object Hallucination in Large Vision-Language Models
- Authors
- Yifan Li, Yifan Du, Kun Zhou, Jinpeng Wang, and 2 others
- Published
- arXiv
- 2305.10355
Abstract
Inspired by the superior language abilities of large language models (LLM), large vision-language models (LVLM) have been recently explored by integrating powerful LLMs for improving the performance on complex multimodal tasks. Despite the promising progress on LVLMs, we find that LVLMs suffer from the hallucination problem, i.e. they tend to generate objects that are inconsistent with the target images in the descriptions. To investigate it, this work presents the first systematic study on object hallucination of LVLMs. We conduct the evaluation experiments on several representative LVLMs, and show that they mostly suffer from severe object hallucination issue. We further discuss that the visual instructions may influence the hallucination, and find that: objects that frequently occur in the visual instructions or co-occur with the image objects, are obviously prone to be hallucinated by LVLMs. Besides, we find that existing evaluation methods might be affected by the input instructions and generation styles of LVLMs. Thus, we further design an improved evaluation method for object hallucination by proposing a polling-based query method called POPE. Experiment results demonstrate that our POPE can evaluate the object hallucination in a more stable and flexible way. Our codes and data are publicly available at https://github.com/RUCAIBox/POPE.
Phi-3.5-vision-instruct leads with 86.1%, followed by Phi-4-multimodal-instruct at 85.6%.
Progress Over Time
Interactive timeline showing model performance evolution on POPE
POPE Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | Microsoft | 4B | — | — | ||
| 2 | Microsoft | 6B | — | — |
FAQ
Common questions about POPE.
More evaluations to explore
Related benchmarks in the same category
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions
MMMU (Massive Multi-discipline Multimodal Understanding) is a benchmark designed to evaluate multimodal models on college-level subject knowledge and deliberate reasoning. Contains 11.5K meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines: Art & Design, Business, Science, Health & Medicine, Humanities & Social Science, and Tech & Engineering across 30 subjects and 183 subfields.
A more robust multi-discipline multimodal understanding benchmark that enhances MMMU through a three-step process: filtering text-only answerable questions, augmenting candidate options, and introducing vision-only input settings. Achieves significantly lower model performance (16.8-26.9%) compared to original MMMU, providing more rigorous evaluation that closely mimics real-world scenarios.
CharXiv-R is the reasoning component of the CharXiv benchmark, focusing on complex reasoning questions that require synthesizing information across visual chart elements. It evaluates multimodal large language models on their ability to understand and reason about scientific charts from arXiv papers through various reasoning tasks.
MathVista evaluates mathematical reasoning of foundation models in visual contexts. It consists of 6,141 examples derived from 28 existing multimodal datasets and 3 newly created datasets (IQTest, FunctionQA, and PaperQA), combining challenges from diverse mathematical and visual tasks to assess models' ability to understand complex figures and perform rigorous reasoning.
AI2D is a dataset of 4,903 illustrative diagrams from grade school natural sciences (such as food webs, human physiology, and life cycles) with over 15,000 multiple choice questions and answers. The benchmark evaluates diagram understanding and visual reasoning capabilities, requiring models to interpret diagrammatic elements, relationships, and structure to answer questions about scientific concepts represented in visual form.