Text Classification on Dreaddit (full)
82.3AccuracyRLIE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RLIEBackbone=DeepSeek-V3.22025.10 | 82.3 | 82.3 | |
| RLIEBackbone=Qwen3-Next-80B2025.10 | 81.1 | 81.1 | |
| HypoGeniCBackbone=DeepSeek-V3.22025.10 | 80.5 | 80.5 | |
| RLIEBackbone=Qwen3-235B2025.10 | 79.9 | 79.9 | |
| IO RefinementBackbone=DeepSeek-V3.22025.10 | 78.5 | 78.1 | |
| Zero-shot GenBackbone=DeepSeek-V3.22025.10 | 67.2 | 63.9 | |
| Zero-shotBackbone=DeepSeek-V3.22025.10 | 64.7 | 60 | |
| Few-shot (ICL)Backbone=DeepSeek-V3.22025.10 | 63.8 | 58.5 |