Phd Graph Learning @ CNAM Paris
@YKarmim on Twitter
Current LLMs often struggle with cultural references of underrepresented groups, leading to strong biases and hallucinations.
Current LLMs often struggle with cultural references of underrepresented groups, leading to strong biases and hallucinations.
We propose a simple yet effective post-hoc model that learns to capture ambiguity from both images and captions, achieving high accuracy in detecting VLM failures.
We propose a simple yet effective post-hoc model that learns to capture ambiguity from both images and captions, achieving high accuracy in detecting VLM failures.