Training Data Reconstruction on RT
1ROUGE-1UTR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| UTRBatch Size=1, Model Architecture=BERT2026.01 | 1 | 1 | |
| UTRBatch Size=1, Model Architecture=GPT22026.01 | 1 | 1 | |
| UTRBatch Size=1, Model Architecture=Qwen2026.01 | 1 | 1 | |
| DAGERBatch Size=1, Model Architecture=BERT2026.01 | 1 | 1 | |
| DAGERBatch Size=1, Model Architecture=GPT22026.01 | 1 | 1 | |
| UTRBatch Size=2, Model Architecture=BERT2026.01 | 1 | 1 | |
| UTRBatch Size=2, Model Architecture=GPT22026.01 | 1 | 1 | |
| UTRBatch Size=2, Model Architecture=Qwen2026.01 | 1 | 1 | |
| DAGERBatch Size=2, Model Architecture=GPT22026.01 | 1 | 1 | |
| UTRBatch Size=4, Model Architecture=BERT2026.01 | 1 | 1 | |
| UTRBatch Size=4, Model Architecture=GPT22026.01 | 1 | 1 | |
| UTRBatch Size=4, Model Architecture=Qwen2026.01 | 1 | 1 | |
| UTRBatch Size=8, Model Architecture=BERT2026.01 | 1 | 1 | |
| UTRBatch Size=8, Model Architecture=GPT22026.01 | 1 | 1 | |
| UTRBatch Size=8, Model Architecture=Qwen2026.01 | 1 | 1 | |
| DAGERBatch Size=2, Model Architecture=BERT2026.01 | 0.981 | 0.965 | |
| DAGERBatch Size=4, Model Architecture=BERT2026.01 | 0.668 | 0.501 | |
| LAMPBatch Size=1, Model Architecture=GPT22026.01 | 0.647 | 0.165 | |
| LAMPBatch Size=2, Model Architecture=GPT22026.01 | 0.464 | 0.076 | |
| DAGERBatch Size=8, Model Architecture=BERT2026.01 | 0.371 | 0.114 | |
| LAMPBatch Size=4, Model Architecture=GPT22026.01 | 0.351 | 0.042 | |
| LAMPBatch Size=1, Model Architecture=BERT2026.01 | 0.314 | 0.093 | |
| DAGERBatch Size=4, Model Architecture=GPT22026.01 | 0.2956 | 0.2165 | |
| LAMPBatch Size=8, Model Architecture=GPT22026.01 | 0.273 | 0.02 | |
| DAGERBatch Size=8, Model Architecture=GPT22026.01 | 0.2037 | 0.1306 | |
| LAMPBatch Size=2, Model Architecture=BERT2026.01 | 0.112 | 0.009 | |
| LAMPBatch Size=8, Model Architecture=BERT2026.01 | 0.068 | 0.003 | |
| LAMPBatch Size=4, Model Architecture=BERT2026.01 | 0.063 | 0.009 | |
| DAGERBatch Size=1, Model Architecture=Qwen2026.01 | 0 | 0 | |
| DAGERBatch Size=2, Model Architecture=Qwen2026.01 | 0 | 0 | |
| DAGERBatch Size=4, Model Architecture=Qwen2026.01 | 0 | 0 | |
| DAGERBatch Size=8, Model Architecture=Qwen2026.01 | 0 | 0 |