Correctness simulation accuracy on DBE-KT22 20% historical interaction logs
69.9AccuracyEdu-Theater
Evaluation Results
| Method | Links | |
|---|---|---|
| Edu-TheaterBase LLM=GPT-4.1-mini2026.06 | 69.9 | |
| Agent4EduBase LLM=GPT-4.1-mini2026.06 | 69.32 | |
| EERNN2026.06 | 68.53 | |
| EduAgentBase LLM=GPT-4.1-mini2026.06 | 68.42 | |
| SAKT2026.06 | 67.88 | |
| Edu-TheaterBase LLM=Llama3-8b2026.06 | 66.33 | |
| Agent4EduBase LLM=Llama3-8b2026.06 | 66.02 | |
| EduAgentBase LLM=Llama3-8b2026.06 | 65.67 | |
| DAISIM2026.06 | 65.25 | |
| DKVMN2026.06 | 60.14 | |
| KES2026.06 | 55.31 |