Argumentation Quality Scoring on STEM Argumentation scoring dataset
68.7Exact MatchGPT-5.2
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-5.2Prompt type=Calibrated Prompt2026.04 | 68.7 | 89.9 | 0.44 | 0.03 | 0.71 | 0.708 | |
| GPT-5.1Prompt type=Calibrated Prompt2026.04 | 67.6 | 89.9 | 0.45 | 0.06 | 0.693 | 0.687 | |
| GPT-4o-miniPrompt type=Calibrated Prompt2026.04 | 50.3 | 91.1 | 0.6 | 0.11 | 0.69 | 0.686 |