Straighten Rope on SoftGym
100Success RateGT Dense (Oracle)
Evaluation Results
| Method | Links | |
|---|---|---|
| GT Dense (Oracle)Reward Signal Type=Hand-crafted, incremental reward, RL Algorithm=SAC2026.07 | 100 | |
| GT Sparse (Oracle)Reward Signal Type=Binary success/failure signals, RL Algorithm=SAC2026.07 | 100 | |
| CLIPReward Signal Type=Cosine similarity, RL Algorithm=SAC2026.07 | 100 | |
| RL-VLM-FReward Signal Type=Preference-based reward, RL Algorithm=SAC2026.07 | 100 | |
| VLM-AR3LReward Signal Type=VLM-AR3L reward model, RL Algorithm=SAC2026.07 | 100 |