Behavioral Response Classification on ADAPT N=122 (test)
64Micro-averaged F1Majority Vote
Evaluation Results
| Method | Links | |
|---|---|---|
| Majority VotePrompt Level=Level 4 (Full Prompt)2026.05 | 64 | |
| DebatePrompt Level=Level 4 (Full Prompt)2026.05 | 63.4 | |
| GLM-4.6Prompt Level=Level 4 (Full Prompt)2026.05 | 61.5 | |
| Direct JudgePrompt Level=Level 4 (Full Prompt)2026.05 | 58.4 | |
| GPT-OSSPrompt Level=Level 4 (Full Prompt)2026.05 | 58.2 | |
| Qwen-3Prompt Level=Level 4 (Full Prompt)2026.05 | 58.2 | |
| Majority VotePrompt Level=Level 1 (Category)2026.05 | 54.5 | |
| Qwen-3Prompt Level=Level 1 (Category)2026.05 | 51.3 | |
| GLM-4.6Prompt Level=Level 1 (Category)2026.05 | 50.7 | |
| GPT-OSSPrompt Level=Level 1 (Category)2026.05 | 45.6 |