Independent Tagging on Amazon, Book, Teams, and Sushi
0.982AccuracyCAST
Evaluation Results
| Method | Links | |
|---|---|---|
| CASTBackbone=GPT-5.22026.01 | 0.982 | |
| CASTBackbone=Gemini-3-Flash2026.01 | 0.968 | |
| Few-shotBackbone=DeepSeek-V3.22026.01 | 0.966 | |
| Self-ConsistencyBackbone=GPT-5.22026.01 | 0.962 | |
| TbSBackbone=GPT-5.22026.01 | 0.96 | |
| BaselineBackbone=Gemini-3-Flash2026.01 | 0.96 | |
| CASTBackbone=DeepSeek-V3.22026.01 | 0.956 | |
| BaselineBackbone=GPT-5.22026.01 | 0.95 | |
| APBackbone=DeepSeek-V3.22026.01 | 0.95 | |
| APBackbone=Gemini-3-Flash2026.01 | 0.95 | |
| TbSBackbone=Gemini-3-Flash2026.01 | 0.941 | |
| Few-shotBackbone=GPT-5.22026.01 | 0.931 | |
| APBackbone=GPT-5.22026.01 | 0.93 | |
| Self-ConsistencyBackbone=DeepSeek-V3.22026.01 | 0.93 | |
| TbSBackbone=DeepSeek-V3.22026.01 | 0.93 | |
| Self-ConsistencyBackbone=Gemini-3-Flash2026.01 | 0.93 | |
| BaselineBackbone=DeepSeek-V3.22026.01 | 0.927 | |
| Few-shotBackbone=Gemini-3-Flash2026.01 | 0.92 |