ADAPT reduces MLLM hallucinations 40-60% by aligning cross-attention dynamics via visual anchors, supervised inference, and preference tuning while preserving general capabilities.
hub Mixed citations
Rlaif-v: Aligning mllms through open-source 11 ai feedback for super gpt-4v trustworthiness
Mixed citation behavior. Most common role is method (50%).
16
Pith papers citing it
Method
50% of classified citations
hub tools
citation-role summary
method 3
background 2
dataset 1
citation-polarity summary
representative citing papers
OPPO is an evidence-aware preference optimization objective that contrasts faithful responses under varying visual evidence strengths to reduce hallucinations in MLLMs.
ViPSy constructs policy-aligned and visually grounded preference pairs for VLMs via visual cues from image variants, yielding SOTA hallucination reductions of 35.7% on AMBER and 24.5% on Object HalBench.
LVLMs show vocabulary hijacking by inert tokens that decode to hijacking anchors; HABI locates them, NHAR finds resilient heads, and HAVAE boosts those heads to cut hallucinations.
CAST reduces object hallucination in LVLMs by 6.03% on average across five models and five benchmarks by identifying caption-sensitive attention heads and applying optimized steering directions to their outputs, with negligible added inference cost.
S2H-DPO generates hierarchical prompt-driven preference pairs to improve multi-image reasoning in VLMs while keeping single-image performance intact.
LLM Orchestration integrates modality experts via an LLM controller, cross-modal memory, and interaction layer to enable multimodal input-output without gradient-based training.
SENTINEL reduces MLLM object hallucinations by over 90% via sentence-level early intervention with detector-bootstrapped preference data and C-DPO loss, outperforming prior SOTA on hallucination and capability benchmarks.
Visual-RFT applies reinforcement learning with verifiable perception rewards to improve large vision-language models on fine-grained classification, few-shot detection, and grounding tasks.
InternVL 2.5 is the first open-source MLLM to surpass 70% on the MMMU benchmark via model, data, and test-time scaling, with a 3.7-point gain from chain-of-thought reasoning.
Mixed Preference Optimization with the MMPR dataset boosts multimodal CoT reasoning, lifting InternVL2-8B to 67.0 accuracy on MathVista (+8.7 points) and matching the 76B model.
PStar adaptively selects pseudocode-based reasoning strategies via a Difficulty Feature Vector to reduce hallucinations in vision-language models, reporting SOTA results on POPE and MMStar benchmarks.
MiniCPM-Llama3-V 2.5 delivers GPT-4V-level multimodal performance on phones through architecture, pretraining, and alignment optimizations.
The survey organizes causes of hallucinations in MLLMs, reviews evaluation benchmarks and metrics, and outlines mitigation approaches plus open questions.
Vision-EKIPL injects high-quality actions from external models into RL training to expand exploration and raise the reasoning ceiling of MLLMs, reporting up to 5% gains on the Reason-RFT-CoT benchmark.
A roadmap that defines architectural nativity for multimodal models and categorizes them into Multi-to-Text, Multi-to-Target, and Multi-to-Multi types while outlining an industrial pipeline toward unified transformer-based native multimodal modeling.
citing papers explorer
-
Pseudocode-Guided Structured Reasoning for Automating Reliable Inference in Vision-Language Models
PStar adaptively selects pseudocode-based reasoning strategies via a Difficulty Feature Vector to reduce hallucinations in vision-language models, reporting SOTA results on POPE and MMStar benchmarks.