Overall Vision-Language Performance on Vision-Language Tasks Aggregate
86.8Targeted ASRCroPA
Evaluation Results
| Method | Links | |
|---|---|---|
| CroPATarget Prompt=metaphor, VLM Backbone=Flamingo2025.06 | 86.8 | |
| CroPATarget Prompt=I am sorry, VLM Backbone=Flamingo2025.06 | 79.85 | |
| Multi-PTarget Prompt=metaphor, VLM Backbone=Flamingo2025.06 | 79.33 | |
| CroPATarget Prompt=unknown, VLM Backbone=Flamingo2025.06 | 77.08 | |
| Multi-PTarget Prompt=I am sorry, VLM Backbone=Flamingo2025.06 | 73.18 | |
| CroPATarget Prompt=very good, VLM Backbone=Flamingo2025.06 | 70.2 | |
| CroPATarget Prompt=too late, VLM Backbone=Flamingo2025.06 | 69.2 | |
| Multi-PTarget Prompt=very good, VLM Backbone=Flamingo2025.06 | 61.75 | |
| Multi-PTarget Prompt=unknown, VLM Backbone=Flamingo2025.06 | 60.95 | |
| Multi-PTarget Prompt=too late, VLM Backbone=Flamingo2025.06 | 59.18 | |
| CroPATarget Prompt=not sure, VLM Backbone=Flamingo2025.06 | 50.33 | |
| Single-PTarget Prompt=metaphor, VLM Backbone=Flamingo2025.06 | 39.73 | |
| Multi-PTarget Prompt=not sure, VLM Backbone=Flamingo2025.06 | 39.3 | |
| Single-PTarget Prompt=I am sorry, VLM Backbone=Flamingo2025.06 | 36.4 | |
| Single-PTarget Prompt=very good, VLM Backbone=Flamingo2025.06 | 31 | |
| Single-PTarget Prompt=unknown, VLM Backbone=Flamingo2025.06 | 28.65 | |
| Single-PTarget Prompt=too late, VLM Backbone=Flamingo2025.06 | 24.55 | |
| Single-PTarget Prompt=not sure, VLM Backbone=Flamingo2025.06 | 23.08 |