Explanation Generation
Papers
Showing 231–235 of 235 papers
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | VLIS (Lynx) | Accuracy | 80 | — | Unverified |
| 2 | VLIS (LLaVA) | Accuracy | 73 | — | Unverified |
| 3 | Ground-truth Caption -> GPT3 (Oracle) | Human (%) | 68 | — | Unverified |
| 4 | Predicted Caption -> GPT3 | Human (%) | 33 | — | Unverified |
| 5 | BLIP2 FlanT5-XXL (Fine-tuned) | Human (%) | 27 | — | Unverified |
| 6 | BLIP2 FlanT5-XL (Fine-tuned) | Human (%) | 15 | — | Unverified |
| 7 | BLIP2 FlanT5-XXL (Zero-shot) | Human (%) | 0 | — | Unverified |