Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Goal reconstruction on OneStop
Loading...
54.8
Accuracy (Question Word)
DalEye-LLaVA
31.816
37.783
43.75
49.717
May 4, 2025
Accuracy (Question Word)
Accuracy (Question Category)
BLEU
Updated 3mo ago
Evaluation Results
Method
Method
Links
Accuracy (Question Word)
Accuracy (Question Category)
BLEU
DalEye-LLaVA
Input Modality=Eye-tra...
2025.05
54.8
46.6
23.1
DalEye-Llama
Input Modality=Eye-tra...
2025.05
54.5
45
25.3
Text-only Llama 3.1
Input Modality=Text-only
2025.05
54
45.4
23.6
Text-only LLaVA OneVision
Input Modality=Text-only
2025.05
49.9
42.4
21
DalEye-GPT
Input Modality=Eye-tra...
2025.05
49.5
48.8
25.7
Text-only GPT-4o-mini
Input Modality=Text-only
2025.05
45
43
24
Incorrect Human (different critical span)
Input Modality=Human P...
2025.05
40.7
41
21.9
Gemini few-shot
Prompting Strategy=few...
2025.05
39.5
41.8
27.2
Gemini zero-shot
Prompting Strategy=zer...
2025.05
37.8
33.5
24.7
Incorrect Human (same critical span)
Input Modality=Human P...
2025.05
37
34.9
29.8
Arbitrary Gemini 3
Prompting Strategy=Arb...
2025.05
32.7
36.7
19.4
Feedback
Search any
task
Search any
task