Silhouette-to-animal generation on Silhouette-to-animal n=40 (test)
50.98Animal PlausibilityGPT
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GPTmode=zero-drift prompting2025.06 | 50.98 | — | 56.86 | 17.65 | — | 9.8 | 52.94 | |
| Geminimode=zero-drift prompting2025.06 | 27.45 | — | 23.53 | 17.65 | — | 15.69 | 23.53 | |
| Grokmode=zero-drift prompting2025.06 | 13.73 | — | 9.8 | 13.73 | — | 7.84 | 11.76 | |
| Shape2Animal2025.06 | 7.84 | — | 9.8 | 50.98 | — | 66.67 | 11.76 | |
| GrokEvaluator=Claude Sonnet 4.62025.06 | 4.88 | 3.62 | 3.98 | 3.48 | 3.99 | — | — | |
| GeminiEvaluator=Claude Sonnet 4.62025.06 | 4.85 | 3.67 | 4.03 | 3.67 | 4.06 | — | — | |
| GPTEvaluator=Claude Sonnet 4.62025.06 | 4.8 | 3.92 | 4.12 | 3.92 | 4.19 | — | — | |
| Shape2AnimalEvaluator=Claude Sonnet 4.62025.06 | 4.55 | 3.05 | 3.33 | 3.45 | 3.59 | — | — |