Nearly all models show prompt-induced hallucinations close to or exceeding the number of catastrophic hallucinations when presented with counterfactual questions, indicating that Large Vision-Language Models (LVLMs) are highly vulnerable to such attacks.

Authors

• Person: Not available Organization: arXiv
  Detecting and Evaluating Medical Hallucinations in Large Vision ...

Sources

• Detecting and Evaluating Medical Hallucinations in Large Vision ... arxiv.org arXiv via serper

Referenced by nodes (2)

• Large Vision-Language Models concept
• prompt-induced hallucination concept