Ambiguity Detection on AmbiEnt
74.6F1 ScoreShaQ (Max Span)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ShaQ (Max Span)Backbone=GPT-5.4-mini2026.05 | 74.6 | 77.8 | 75.19 | |
| ShaQBackbone=GPT-5.4-mini2026.05 | 73.84 | 77.68 | 75.71 | |
| LOOBackbone=GPT-5.4-mini2026.05 | 66.07 | 49.04 | 57 | |
| LOO (Max Span)Backbone=GPT-5.4-mini2026.05 | 66.07 | 52.66 | 59.69 | |
| Ask4Conf-DBackbone=GPT-5.4-mini2026.05 | 65.43 | 50.82 | 52.54 | |
| ICEBackbone=GPT-5.4-mini2026.05 | 57.33 | 55.4 | 53.21 | |
| Deep EnsemblesBackbone=GPT-5.4-mini2026.05 | 56.37 | 58.96 | 61.49 | |
| Sample DiversityBackbone=GPT-5.4-mini2026.05 | 48.33 | 59.23 | 57.48 | |
| Self-ConsistencyBackbone=GPT-5.4-mini2026.05 | 47.93 | 59.28 | 59.46 | |
| Sample RepetitionBackbone=GPT-5.4-mini2026.05 | 43.69 | 56.67 | 57.78 | |
| Semantic EntropyBackbone=GPT-5.4-mini2026.05 | 42.2 | 57.56 | 54.36 |