E-commerce Environment Navigation Calibration on WebShop
0.185ECE (End-State)Forward UQ (UAM)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Forward UQ (UAM)module=UAM-only2026.01 | 0.185 | 0.245 | 0.221 | 0.228 | 0.138 | 0.268 | |
| Dual-Process (AUQ)module=AUQ, Best-of-N=32026.01 | 0.188 | 0.242 | 0.215 | 0.236 | 0.21 | 0.235 | |
| Inverse UQ (UAR)module=UAR-only, Best-of-N=32026.01 | 0.209 | 0.231 | 0.242 | 0.222 | 0.262 | 0.248 | |
| CoT-SC2026.01 | 0.225 | 0.263 | 0.269 | 0.253 | 0.242 | 0.25 | |
| Self-Reflection2026.01 | 0.31 | 0.278 | 0.295 | 0.288 | 0.298 | 0.276 | |
| ReAct2026.01 | 0.335 | 0.288 | 0.342 | 0.305 | 0.329 | 0.298 | |
| Reflexion2026.01 | 0.352 | 0.275 | 0.334 | 0.296 | 0.272 | 0.284 |