Student Misconception Classification on MAP-Charting
95.85MAP@3Qwen-3-4B + Ours
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen-3-4B + OursEvaluation Protocol=Two-Stage Distillation2026.05 | 95.85 | 95.87 | 91.98 | 49.96 | |
| Gemma-2-9B + OursEvaluation Protocol=Two-Stage Distillation2026.05 | 95.6 | 95.62 | 91.48 | 49.95 | |
| Llama-3.1-8B + OursEvaluation Protocol=Two-Stage Distillation2026.05 | 95.53 | 95.55 | 91.34 | 49.95 | |
| Qwen-2.5-72BEvaluation Protocol=Fine-tuned2026.05 | 94.97 | 95.01 | 90.14 | 49.93 | |
| Qwen-3-4BEvaluation Protocol=Fine-tuned2026.05 | 94.72 | 94.75 | 89.87 | 49.92 | |
| Llama-3.1-8BEvaluation Protocol=Fine-tuned2026.05 | 94.53 | 94.56 | 89.54 | 49.9 | |
| Gemma-2-9BEvaluation Protocol=Fine-tuned2026.05 | 94.39 | 94.42 | 89.19 | 49.92 | |
| GPT-5Evaluation Protocol=Prompting2026.05 | 81.37 | 81.45 | 72.25 | 46.26 | |
| Claude-4-SonnetEvaluation Protocol=Prompting2026.05 | 78.33 | 78.41 | 69.14 | 45.79 | |
| DeepSeek-V3Evaluation Protocol=Prompting2026.05 | 76.65 | 76.73 | 66.01 | 45.05 | |
| GPT-OSS-120BEvaluation Protocol=Prompting2026.05 | 76.61 | 76.69 | 67.94 | 43.75 | |
| Qwen-2.5-72BEvaluation Protocol=Prompting2026.05 | 72.85 | 72.93 | 62.22 | 43.28 |