Dimensional Aspect-Based Sentiment Triplet Extraction on DimASTE Combined Lang-Domain Sets
54.42cF1GPT-OSS 120B
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-OSS 120BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Restaurant2026.03 | 54.42 | |
| Llama 3.3 70BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Restaurant2026.03 | 54.18 | |
| Kimi K2 ThinkingEvaluation Protocol=Zero-shot, Language=ENG, Domain=Restaurant2026.03 | 51.01 | |
| GPT-5 miniEvaluation Protocol=One-shot, Language=ENG, Domain=Restaurant2026.03 | 50.34 | |
| GPT-5 miniEvaluation Protocol=Zero-shot, Language=ENG, Domain=Restaurant2026.03 | 49.93 | |
| Kimi K2 ThinkingEvaluation Protocol=One-shot, Language=ENG, Domain=Restaurant2026.03 | 49.2 | |
| GPT-5 miniEvaluation Protocol=One-shot, Language=ENG, Domain=Laptop2026.03 | 48.74 | |
| Llama 3.3 70BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Laptop2026.03 | 46.64 | |
| Kimi K2 ThinkingEvaluation Protocol=Zero-shot, Language=ENG, Domain=Laptop2026.03 | 45.19 | |
| GPT-OSS 120BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Laptop2026.03 | 45.15 | |
| GPT-5 miniEvaluation Protocol=Zero-shot, Language=ENG, Domain=Laptop2026.03 | 44.91 | |
| Qwen 3 14BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Restaurant2026.03 | 44.83 | |
| Kimi K2 ThinkingEvaluation Protocol=One-shot, Language=ENG, Domain=Laptop2026.03 | 44.24 | |
| Qwen 3 14BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Laptop2026.03 | 38.27 | |
| Mistral 3 14BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Restaurant2026.03 | 29.3 | |
| Mistral 3 14BEvaluation Protocol=Supervised Fine-Tuning, Language=ENG, Domain=Laptop2026.03 | 27.36 |