Codebook Generation and Revision on UCSB sub-theme level
61.1RecallGPT-5 (Data)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-5 (Data)Perspective=Data-revised2026.05 | 61.1 | 68.6 | 34.3 | 87.5 | |
| Qwen3.5-P (Applied)Perspective=Applied-revised2026.05 | 61.1 | 95.2 | 21.4 | 23.2 | |
| GPT-5Perspective=Original2026.05 | 55.6 | 70.4 | 32.4 | 87.6 | |
| GPT-5 (Theory)Perspective=Theory-revised2026.05 | 55.6 | 68.5 | 32.9 | 87.6 | |
| GPT-5 (Applied)Perspective=Applied-revised2026.05 | 55.6 | 66.2 | 28.4 | 87.4 | |
| DS-V3.2Perspective=Original2026.05 | 55.6 | 82.1 | 21.4 | 39.1 | |
| Qwen3.5-P (Data)Perspective=Data-revised2026.05 | 55.6 | 83.3 | 25 | 23.2 | |
| DS-V3.2 (Applied)Perspective=Applied-revised2026.05 | 50 | 84.1 | 22.7 | 29.3 | |
| DS-V3.2 (Data)Perspective=Data-revised2026.05 | 50 | 80 | 20 | 39.1 | |
| Qwen3.5-PPerspective=Original2026.05 | 50 | 94.2 | 21.2 | 23.3 | |
| Qwen3.5-P (Theory)Perspective=Theory-revised2026.05 | 50 | 91.1 | 15.6 | 23.3 | |
| DS-V3.2 (Theory)Perspective=Theory-revised2026.05 | 44.4 | 86.7 | 20 | 39.1 |