Response-Use classification on RelianceScope student logs (chat, code-edit, and copy)
0.76F1 Passivegemini-1.5-pro-preview
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| gemini-1.5-pro-previewPrompting=Zero-Shot2026.02 | 0.76 | 0.633 | |
| gemini-1.5-pro-previewPrompting=3-Shot2026.02 | 0.722 | 0.58 | |
| gemini-1.5-pro-previewPrompting=9-Shot2026.02 | 0.704 | 0.58 | |
| gemini-1.5-pro-previewPrompting=9-Shot+CoT2026.02 | 0.678 | 0.58 | |
| gemini-1.5-flash-previewPrompting=9-Shot2026.02 | 0.667 | 0.533 | |
| gemini-1.5-flash-previewPrompting=9-Shot+CoT2026.02 | 0.667 | 0.58 | |
| qwen2.5:8bPrompting=9-Shot2026.02 | 0.646 | 0.453 | |
| qwen2.5:8bPrompting=Zero-Shot2026.02 | 0.635 | 0.427 | |
| qwen2.5:8bPrompting=3-Shot2026.02 | 0.628 | 0.427 | |
| gemini-1.5-flash-previewPrompting=3-Shot2026.02 | 0.593 | 0.473 | |
| qwen2.5:8bPrompting=9-Shot+CoT2026.02 | 0.593 | 0.46 | |
| gemini-1.5-flash-previewPrompting=Zero-Shot2026.02 | 0.585 | 0.487 |