Persona Discrimination on NeurIPS Cross-conference
0.418Persona Separability (Δ)TADDLE
Evaluation Results
| Method | Links | |
|---|---|---|
| TADDLEFamily=Ours2026.05 | 0.418 | |
| Qwen3-30BFamily=Few-shot LLM2026.05 | 0.393 | |
| Qwen3.5-PlusFamily=Few-shot LLM2026.05 | 0.387 | |
| RottenReviews (Llama-3-8B FT)Family=Review-quality detector, Backbone=Llama-3-8B, Setting=Fine-tuned2026.05 | 0.367 | |
| GLM-4.7Family=Few-shot LLM2026.05 | 0.366 | |
| ReviewGuard (Llama-3.1-8B)Family=Review-quality detector, Backbone=Llama-3.1-8B2026.05 | 0.31 | |
| ReviewGuard (Qwen3-8B)Family=Review-quality detector, Backbone=Qwen3-8B2026.05 | 0.283 | |
| Longformer-base-4096Family=PLM2026.05 | 0.277 | |
| BigBird-RoBERTa-baseFamily=PLM2026.05 | 0.242 | |
| DeepSeek-V3.2Family=Few-shot LLM2026.05 | 0.235 | |
| RottenReviews (XGBoost)Family=Review-quality detector, Classifier=XGBoost2026.05 | 0.216 | |
| RottenReviews (Random Forest)Family=Review-quality detector, Classifier=Random Forest2026.05 | 0.206 | |
| BERT-base-uncasedFamily=PLM2026.05 | 0.201 | |
| RottenReviews (LR)Family=Review-quality detector, Classifier=LR2026.05 | 0.057 | |
| RottenReviews (MLP)Family=Review-quality detector, Classifier=MLP2026.05 | 0.003 | |
| RottenReviews (Qwen3-8B 0-shot)Family=Review-quality detector, Backbone=Qwen3-8B, Setting=0-shot2026.05 | 0 |