How Easy is It to Fool Your Multimodal LLMs? An Empirical Analysis on Deceptive Prompts

• URL: http://arxiv.org/abs/2402.13220v2
• Date: Tue, 23 Jul 2024 07:02:30 GMT
• Title: How Easy is It to Fool Your Multimodal LLMs? An Empirical Analysis on Deceptive Prompts
• Authors: Yusu Qian, Haotian Zhang, Yinfei Yang, Zhe Gan,
• Abstract summary: We present MAD-Bench, a benchmark that contains 1000 test samples divided into 5 categories, such as non-existent objects, count of objects, and spatial relationship. We provide a comprehensive analysis of popular MLLMs, ranging from GPT-4v, Reka, Gemini-Pro, to open-sourced models, such as LLaVA-NeXT and MiniCPM-Llama3. While GPT-4o achieves 82.82% accuracy on MAD-Bench, the accuracy of any other model in our experiments ranges from 9% to 50%.
• Score: 54.07541591018305
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The remarkable advancements in Multimodal Large Language Models (MLLMs) have not rendered them immune to challenges, particularly in the context of handling deceptive information in prompts, thus producing hallucinated responses under such conditions. To quantitatively assess this vulnerability, we present MAD-Bench, a carefully curated benchmark that contains 1000 test samples divided into 5 categories, such as non-existent objects, count of objects, and spatial relationship. We provide a comprehensive analysis of popular MLLMs, ranging from GPT-4v, Reka, Gemini-Pro, to open-sourced models, such as LLaVA-NeXT and MiniCPM-Llama3. Empirically, we observe significant performance gaps between GPT-4o and other models; and previous robust instruction-tuned models are not effective on this new benchmark. While GPT-4o achieves 82.82% accuracy on MAD-Bench, the accuracy of any other model in our experiments ranges from 9% to 50%. We further propose a remedy that adds an additional paragraph to the deceptive prompts to encourage models to think twice before answering the question. Surprisingly, this simple method can even double the accuracy; however, the absolute numbers are still too low to be satisfactory. We hope MAD-Bench can serve as a valuable benchmark to stimulate further research to enhance model resilience against deceptive prompts.

Related papers

• SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal Behaviors [64.9938658716425]
  Existing evaluations of large language models' (LLMs) ability to recognize and reject unsafe user requests face three limitations. First, existing methods often use coarse-grained of unsafe topics, and are over-representing some fine-grained topics. Second, linguistic characteristics and formatting of prompts are often overlooked, like different languages, dialects, and more -- which are only implicitly considered in many evaluations. Third, existing evaluations rely on large LLMs for evaluation, which can be expensive.
  arXiv  Detail & Related papers  (2024-06-20T17:56:07Z)
• UBENCH: Benchmarking Uncertainty in Large Language Models with Multiple Choice Questions [10.28688988951815]
  UBENCH is a benchmark for evaluating large language models. It includes 3,978 multiple-choice questions covering knowledge, language, understanding, and reasoning abilities. We also evaluate the reliability of 15 popular LLMs, finding GLM4 to be the most outstanding.
  arXiv  Detail & Related papers  (2024-06-18T16:50:38Z)
• RLAIF-V: Aligning MLLMs through Open-Source AI Feedback for Super GPT-4V Trustworthiness [94.03511733306296]
  We introduce RLAIF-V, a framework that aligns MLLMs in a fully open-source paradigm for super GPT-4V trustworthiness. RLAIF-V maximally exploits the open-source feedback from two perspectives, including high-quality feedback data and online feedback learning algorithm. Experiments show that RLAIF-V substantially enhances the trustworthiness of models without sacrificing performance on other tasks.
  arXiv  Detail & Related papers  (2024-05-27T14:37:01Z)
• Language Models can Evaluate Themselves via Probability Discrepancy [38.54454263880133]
  We propose a new self-evaluation method ProbDiff for assessing the efficacy of various Large Language Models (LLMs) It uniquely utilizes the LLMs being tested to compute the probability discrepancy between the initial response and its revised versions. Our findings reveal that ProbDiff achieves results on par with those obtained from evaluations based on GPT-4.
  arXiv  Detail & Related papers  (2024-05-17T03:50:28Z)
• Are We on the Right Way for Evaluating Large Vision-Language Models? [92.5761176224556]
  Large vision-language models (LVLMs) have recently achieved rapid progress, sparking numerous studies to evaluate their multi-modal capabilities. We identify two primary issues: Visual content is unnecessary for many samples and intentional data leakage exists. We present MMStar, an elite vision-indispensable multi-modal benchmark comprising 1,500 samples meticulously selected by humans.
  arXiv  Detail & Related papers  (2024-03-29T17:59:34Z)
• Multimodal Large Language Models to Support Real-World Fact-Checking [80.41047725487645]
  Multimodal large language models (MLLMs) carry the potential to support humans in processing vast amounts of information. While MLLMs are already being used as a fact-checking tool, their abilities and limitations in this regard are understudied. We propose a framework for systematically assessing the capacity of current multimodal models to facilitate real-world fact-checking.
  arXiv  Detail & Related papers  (2024-03-06T11:32:41Z)
• Enhancing Multimodal Large Language Models with Vision Detection Models: An Empirical Study [32.57246173437492]
  This paper presents an empirical study on enhancing MLLMs with state-of-the-art (SOTA) object detection and Optical Character Recognition (OCR) models to improve fine-grained understanding and reduce hallucination in responses. We conduct systematic and extensive experiments with representative models such as LLaVA-1.5, DINO, PaddleOCRv2, and Grounding DINO. Notably, the enhanced LLaVA-1.5 outperforms its original 7B/13B models on all 10 benchmarks, achieving an improvement of up to 12.5% on the normalized average score.
  arXiv  Detail & Related papers  (2024-01-31T16:38:32Z)
• Mixed Distillation Helps Smaller Language Model Better Reasoning [27.934081882868902]
  We introduce Mixed Distillation (MD) framework, which capitalizes on the strengths of Program of Thought (PoT) and Chain of Thought (CoT) capabilities within large language models (LLMs) Our experimental results show that MD significantly enhances the single-path and multi-path reasoning ability of smaller models in various tasks.
  arXiv  Detail & Related papers  (2023-12-17T14:28:28Z)
• ReEval: Automatic Hallucination Evaluation for Retrieval-Augmented Large Language Models via Transferable Adversarial Attacks [91.55895047448249]
  This paper presents ReEval, an LLM-based framework using prompt chaining to perturb the original evidence for generating new test cases. We implement ReEval using ChatGPT and evaluate the resulting variants of two popular open-domain QA datasets. Our generated data is human-readable and useful to trigger hallucination in large language models.
  arXiv  Detail & Related papers  (2023-10-19T06:37:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.