Semantic Object Disambiguation on GSR-bench Easy 1.0 (test)
97.5Task ProgressRoBoSR-8B-FT
Evaluation results
| Method | Links | |
|---|---|---|
| RoBoSR-8B-FTmode=fine-tuned, demonstration_trajectories=362026.06 | 97.5 | |
| RoBoSR-8Bmode=zero-shot2026.06 | 70 | |
| Qwen3-8Btype=prompting-based2026.06 | 63 | |
| Gemini-2.5-protype=prompting-based2026.06 | 32.3 | |
| GPT-5type=prompting-based2026.06 | 30 | |
| Claude-sonnet-4.5type=prompting-based2026.06 | 28 | |
| DeepSeek-V3type=prompting-based2026.06 | 23 |