Semantic Object Disambiguation on GSR-bench Difficult 1.0 (test)
89.2Task ProgressRoBoSR-8B-FT
Evaluation results
| Method | Links | |
|---|---|---|
| RoBoSR-8B-FTmode=fine-tuned, demonstration_trajectories=362026.06 | 89.2 | |
| RoBoSR-8Bmode=zero-shot2026.06 | 52 | |
| Qwen3-8Btype=prompting-based2026.06 | 41 | |
| Claude-sonnet-4.5type=prompting-based2026.06 | 19 | |
| Gemini-2.5-protype=prompting-based2026.06 | 17.5 | |
| DeepSeek-V3type=prompting-based2026.06 | 15 | |
| GPT-5type=prompting-based2026.06 | 10 |