Personalized Text Generation on LaMP-5 v1 (test)
0.401ROUGE-LAttn-GS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Attn-GSBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.401 | 0.401 | |
| Attn-GSBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.397 | 0.397 | |
| Attn-GSBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.394 | 0.394 | |
| Attn-GSBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.388 | 0.388 | |
| Prompt-GBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.379 | 0.379 | |
| CoTBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.379 | 0.379 | |
| Self-ReflectionBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.379 | 0.379 | |
| Direct SummaryBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.378 | 0.378 | |
| Mark AllBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.378 | 0.378 | |
| Prompt-GBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.378 | 0.378 | |
| Direct SummaryBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.377 | 0.377 | |
| CoTBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.377 | 0.377 | |
| Mark AllBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.376 | 0.376 | |
| Mark AllBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.376 | 0.376 | |
| Prompt-GBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.376 | 0.376 | |
| Self-ReflectionBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.376 | 0.376 | |
| Direct SummaryBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.375 | 0.375 | |
| Random MarkBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.375 | 0.375 | |
| Self-ReflectionBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.375 | 0.375 | |
| CoTBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.374 | 0.374 | |
| Attn-GSBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.374 | 0.374 | |
| Random MarkBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.373 | 0.373 | |
| Mark AllBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.372 | 0.372 | |
| Attn-GSBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.372 | 0.372 | |
| Direct SummaryBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.371 | 0.371 | |
| Random MarkBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.371 | 0.371 | |
| Prompt-GBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.371 | 0.371 | |
| Truncated ContBackbone=GPT-4o-mini, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.37 | 0.37 | |
| Truncated ContBackbone=GPT-4o-mini, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.37 | 0.37 | |
| CoTBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.37 | 0.37 | |
| Self-ReflectionBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.37 | 0.37 | |
| Attn-GSBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.37 | 0.37 | |
| Truncated ContBackbone=GPT-4o-mini, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.368 | 0.368 | |
| Random MarkBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.367 | 0.367 | |
| Attn-GSBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.359 | 0.359 | |
| Prompt-GBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.355 | 0.355 | |
| Direct SummaryBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.354 | 0.354 | |
| Mark AllBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.354 | 0.354 | |
| Prompt-GBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.354 | 0.354 | |
| Truncated ContBackbone=GPT-4o-mini, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.353 | 0.353 | |
| Direct SummaryBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.353 | 0.353 | |
| Direct SummaryBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.352 | 0.352 | |
| Mark AllBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.352 | 0.352 | |
| Mark AllBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.352 | 0.352 | |
| Prompt-GBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.352 | 0.352 | |
| CoTBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.352 | 0.352 | |
| CoTBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.351 | 0.351 | |
| Self-ReflectionBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.351 | 0.351 | |
| Random MarkBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.35 | 0.35 | |
| Self-ReflectionBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.35 | 0.35 | |
| Random MarkBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.349 | 0.349 | |
| CoTBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.349 | 0.349 | |
| Self-ReflectionBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.349 | 0.349 | |
| Random MarkBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.348 | 0.348 | |
| Truncated ContBackbone=Llama-3-8B, Token Count=200, Evaluation Protocol=Zero-shot2026.02 | 0.345 | 0.345 | |
| Truncated ContBackbone=Llama-3-8B, Token Count=150, Evaluation Protocol=Zero-shot2026.02 | 0.344 | 0.344 | |
| Truncated ContBackbone=Llama-3-8B, Token Count=100, Evaluation Protocol=Zero-shot2026.02 | 0.343 | 0.343 | |
| Mark AllBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.339 | 0.339 | |
| Direct SummaryBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.338 | 0.338 | |
| Prompt-GBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.338 | 0.338 | |
| CoTBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.336 | 0.336 | |
| Self-ReflectionBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.335 | 0.335 | |
| Random MarkBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.334 | 0.334 | |
| Truncated ContBackbone=Llama-3-8B, Token Count=50, Evaluation Protocol=Zero-shot2026.02 | 0.33 | 0.33 |