Personalized Text Generation on LaMP-7 (val)
53.5ROUGE-1 ScorePHFCompass
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PHFCompassParadigm=PHF Implements2026.06 | 53.5 | 47.5 | |
| PPlugParadigm=Flat: Aggregation-Based2026.06 | 52.4 | 46.5 | |
| BM25Paradigm=Flat: Retrieval-Based2026.06 | 46.4 | 40.8 | |
| L-FParadigm=PHF Implements2026.06 | 46.3 | 40.6 | |
| DenseParadigm=Flat: Retrieval-Based2026.06 | 46.2 | 40.6 | |
| T-H+L-FParadigm=PHF Implements2026.06 | 45.9 | 40.3 | |
| T-HParadigm=PHF Implements2026.06 | 45.6 | 40 | |
| Zero-ShotBase Model=Qwen2.5-7B-Instruct2026.06 | 44.8 | 39.2 | |
| ROPGParadigm=Flat: Retrieval-Based2026.06 | 44.7 | 39.3 | |
| DiffMeanParadigm=Flat: Aggregation-Based2026.06 | 44.4 | 38.8 | |
| RecencyParadigm=Flat: Retrieval-Based2026.06 | 41.8 | 36.5 | |
| T-H+T-FParadigm=PHF Implements2026.06 | 41.8 | 36.9 | |
| T-FParadigm=PHF Implements2026.06 | 41.4 | 36.5 |