Hal-Eval: A Universal and Fine-grained Hallucination Evaluation Framework for Large Vision Language Models.
Chaoya JiangWei YeMengfan DongHongrui JiaHaiyang XuMing YanJi ZhangShikun ZhangPublished in: CoRR (2024)
Keyphrases
- fine grained
- language model
- evaluation framework
- language modeling
- coarse grained
- evaluation methodology
- evaluation process
- test collection
- n gram
- document retrieval
- information retrieval
- retrieval model
- query expansion
- probabilistic model
- access control
- evaluation metrics
- semantic annotation
- vector space model
- query terms
- evaluation measures
- test set
- pairwise