R-Bench

'R-Bench' is a benchmark dataset used to evaluate the performance of models in detecting hallucinations in outputs generated by large Vision-Language Models (LVLMs).

Papers using R-Bench (6)

SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning2026Native Visual Understanding: Resolving Resolution Dilemmas In Vision-language Models2025More Thinking, Less Seeing? Assessing Amplified Hallucination In Multimodal Reasoning Models2025Ro-bench: Large-scale Robustness Evaluation Of Mllms With Text-driven Counterfactual Videos2025CutPaste&Find: Efficient Multimodal Hallucination Detector with Visual-aid Knowledge Base2025Evaluating and Analyzing Relationship Hallucinations in Large Vision-Language Models2024