Aspect-based Sentiment Quadruple Extraction on LASQ Uyghur
80.9Entity Target F1H2DT
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| H2DT2026.04 | 80.9 | 24.63 | 49.86 | 20.38 | 39.8 | 17.31 | 20.45 | 11.05 | 14.25 | |
| SKEM2026.04 | 80.46 | 27.75 | 50.66 | 23.63 | 40.01 | 18.28 | 19.77 | 13.6 | 16.11 | |
| Meta-WP2026.04 | 79.74 | 24.68 | 50.25 | 20.23 | 38.45 | 17.23 | 23.52 | 9.77 | 13.81 | |
| MRM-STS2026.04 | 77.51 | 26.11 | 47.94 | 21.21 | 39.71 | 16.68 | 16.04 | 11.8 | 13.59 | |
| EC-ACOS2026.04 | 69.32 | 23.81 | 37.55 | 18.21 | 30.82 | 15.59 | 14.29 | 8.006 | 10.31 | |
| Span-ASTE2026.04 | 68.31 | 24.76 | 39.29 | 18.78 | 31.83 | 15.65 | 15.55 | 9.36 | 11.69 | |
| ParaPhrase2026.04 | 62.65 | 20.95 | 40.26 | 16.72 | 31.77 | 13.57 | 18.63 | 7.26 | 10.45 | |
| GPT-5Evaluation Setting=In-context learning (ICL)2026.04 | 42.83 | 16.14 | 25.11 | 11.85 | 24.39 | 12.43 | 14.2 | 8.51 | 10.64 | |
| GPT-5Evaluation Setting=Zero-shot (ZS)2026.04 | 40.79 | 14.41 | 22.83 | 10.49 | 22.17 | 11 | 12.57 | 6.53 | 8.59 | |
| Qwen3-MaxEvaluation Setting=In-context learning (ICL)2026.04 | 34.21 | 15.8 | 24.63 | 12.15 | 23.65 | 11.78 | 12.4 | 8.34 | 9.97 | |
| Qwen3-MaxEvaluation Setting=Zero-shot (ZS)2026.04 | 31.63 | 12.32 | 20.57 | 9.87 | 20.59 | 9.54 | 10.09 | 6.38 | 7.82 |