Unified Hallucination Fuzzing for Multimodal Large Language Models

• URL: http://arxiv.org/abs/2608.07525v1
• Date: Wed, 15 Jul 2026 17:57:37 GMT
• Title: Unified Hallucination Fuzzing for Multimodal Large Language Models
• Abstract summary: Hallucination remains a persistent challenge for Multimodal Large Language Models (MLLMs)<n>Existing evaluations, predominantly based on static benchmarks, suffer from narrow taxonomical coverage and rapid performance saturation.<n>We present a systematic evaluation framework integrating a comprehensive benchmark with self-evolving stress testing.
• Score: 44.4785399881315
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Hallucination remains a persistent challenge for Multimodal Large Language Models (MLLMs), severely limiting their reliability in high-stakes applications. Existing evaluations, predominantly based on static benchmarks, suffer from narrow taxonomical coverage and rapid performance saturation, failing to reflect model robustness in evolving real-world scenarios. To bridge this gap, we present a systematic evaluation framework integrating a comprehensive benchmark with self-evolving stress testing. First, we introduce UniHall, a fine-grained dataset grounded in a unified taxonomy spanning Object, Instruction, and Knowledge dimensions. Second, to address benchmark saturation, we propose Self-Adaptive Multimodal Fuzzing (SAMF), a self-adaptive framework that employs evolutionary mutation strategies to explore the boundaries of model hallucinations. Crucially, to ensure reliable assessment of dynamic inputs, SAMF incorporates a structured metric suite driven by an ensemble of multi-modal oracles. Our extensive experiments reveal that state-of-the-art MLLMs exhibit significant performance degradation under fuzzing compared to conventional settings, exposing a dissociation between reasoning capabilities and factual grounding. Furthermore, we identify a helpfulness-hallucination trade-off, where reinforcement learning alignment inadvertently exacerbates sycophancy in instruction-following tasks. The framework, code and benchmark are available at https://github.com/LanceZPF/EvalHall.

Related papers

• AnE: Pushing the Reasoning Frontier of Multimodal LLMs via Anchor Evolution [61.593935260052795]
  Post-training viaSupervised Fine-Tuning (SFT) and Reinforcement Learning (RL) is crucial for enhancing reasoning in Multimodal Large Language Models (MLLMs)<n>Existing paradigms often reach a performance bottleneck due to the limitations of static data.<n>We propose Anchor Evolution (AnE), a new paradigm that integrates truth-anchored data curation and model evolution.
  arXiv  Detail & Related papers  (2026-05-25T08:26:34Z)
• LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations [53.20772659095155]
  This paper tackles the more challenging setting of IML under training-time incomplete observations.<n>We propose LIMSSR (LLM-Driven Incomplete Multimodal Sequence-to-Score Reasoning), a framework that reformulates this challenge as a conditional sequence reasoning task.
  arXiv  Detail & Related papers  (2026-05-01T06:11:42Z)
• Diversity in Large Language Models under Supervised Fine-Tuning [15.119045051735633]
  Supervised Fine-Tuning (SFT) is essential for aligning Large Language Models (LLMs) with user intent, yet it is believed to suppress generative diversity.<n>In this study, we attribute the decline to two primary drivers: the forgetting of preexisting knowledge and low-frequency patterns within finetuning datasets.<n>Our evaluation confirms at scale that generation narrows after SFT and strengthens the hypothesis explaining this effect.
  arXiv  Detail & Related papers  (2026-04-30T20:20:59Z)
• UniICL: Systematizing Unified Multimodal In-context Learning through a Capability-Oriented Taxonomy [99.95663439922937]
  In-context Learning efficacy is often non-monotonic and highly task-dependent.<n>We introduce a six-level capability-oriented taxonomy that categorizes the functional role of demonstrations.<n>We construct UniICL-760K, a large-scale corpus featuring curated 8-shot In-context Learning episodes.<n>As an architectural intervention, we propose the Context-Adaptive Prototype Modulator.
  arXiv  Detail & Related papers  (2026-03-25T18:09:33Z)
• PENDULUM: A Benchmark for Assessing Sycophancy in Multimodal Large Language Models [43.767942065379366]
  Sycophancy is a tendency of AI models to agree with user input at the expense of factual accuracy or in contradiction of visual evidence.<n>We introduce a comprehensive evaluation benchmark, textitPENDULUM, comprising approximately 2,000 human-curated Visual Question Answering pairs.<n>We observe substantial variability in model robustness and a pronounced susceptibility to sycophantic and hallucinatory behavior.
  arXiv  Detail & Related papers  (2025-12-22T12:49:12Z)
• Beyond Single Models: Mitigating Multimodal Hallucinations via Adaptive Token Ensemble Decoding [41.828387997311474]
  Large Vision-Language Models (LVLMs) have recently achieved impressive results in multimodal tasks such as image captioning and visual question answering.<n>They remain prone to object hallucination -- generating descriptions of nonexistent or misidentified objects.<n>We propose Adaptive Token Ensemble Decoding (ATED), a training-free, token-level ensemble framework that mitigates hallucination by aggregating predictions from multiple LVLMs during inference.
  arXiv  Detail & Related papers  (2025-10-21T06:11:24Z)
• A Unified Virtual Mixture-of-Experts Framework:Enhanced Inference and Hallucination Mitigation in Single-Model System [9.764336669208394]
  Generative models, such as GPT and BERT, have significantly improved performance in tasks like text generation and summarization.<n>However, hallucinations "where models generate non-factual or misleading content" are especially problematic in smaller-scale architectures.<n>We propose a unified Virtual Mixture-of-Experts (MoE) fusion strategy that enhances inference performance and mitigates hallucinations in a single Qwen 1.5 0.5B model.
  arXiv  Detail & Related papers  (2025-04-01T11:38:01Z)
• MMIE: Massive Multimodal Interleaved Comprehension Benchmark for Large Vision-Language Models [71.36392373876505]
  We introduce MMIE, a large-scale benchmark for evaluating interleaved multimodal comprehension and generation in Large Vision-Language Models (LVLMs)<n>MMIE comprises 20K meticulously curated multimodal queries, spanning 3 categories, 12 fields, and 102 subfields, including mathematics, coding, physics, literature, health, and arts.<n>It supports both interleaved inputs and outputs, offering a mix of multiple-choice and open-ended question formats to evaluate diverse competencies.
  arXiv  Detail & Related papers  (2024-10-14T04:15:00Z)
• Sycophancy in Vision-Language Models: A Systematic Analysis and an Inference-Time Mitigation Framework [18.54098084470481]
  We analyze sycophancy across vision-language benchmarks and propose an inference-time mitigation framework.<n>Our framework effectively mitigates sycophancy across all evaluated models, while maintaining performance on neutral prompts.
  arXiv  Detail & Related papers  (2024-08-21T01:03:21Z)
• VALOR-EVAL: Holistic Coverage and Faithfulness Evaluation of Large Vision-Language Models [57.43276586087863]
  Large Vision-Language Models (LVLMs) suffer from hallucination issues, wherein the models generate plausible-sounding but factually incorrect outputs. Existing benchmarks are often limited in scope, focusing mainly on object hallucinations. We introduce a multi-dimensional benchmark covering objects, attributes, and relations, with challenging images selected based on associative biases.
  arXiv  Detail & Related papers  (2024-04-22T04:49:22Z)

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.