Dimensional Aspect Sentiment Regression on JPN Finance
1.0188RMSE_VAGPT-OSS 120B
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-OSS 120BEvaluation Protocol=Supervised Fine-Tuning, Model=GPT-OSS 120B2026.03 | 1.0188 | |
| Kimi K2 ThinkingEvaluation Protocol=One-shot, Model=Kimi K2 Thinking2026.03 | 1.6396 | |
| Qwen 3 14BEvaluation Protocol=Supervised Fine-Tuning, Model=Qwen 3 14B2026.03 | 1.8964 | |
| GPT-5 miniEvaluation Protocol=One-shot, Model=GPT-5 mini2026.03 | 1.9243 | |
| Mistral 3 14BEvaluation Protocol=Supervised Fine-Tuning, Model=Mistral 3 14B2026.03 | 2.07 | |
| Kimi K2 ThinkingEvaluation Protocol=Zero-shot, Model=Kimi K2 Thinking2026.03 | 2.3379 | |
| Llama 3.3 70BEvaluation Protocol=Supervised Fine-Tuning, Model=Llama 3.3 70B2026.03 | 2.4191 | |
| GPT-5 miniEvaluation Protocol=Zero-shot, Model=GPT-5 mini2026.03 | 2.676 |