PerLTQA
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Membership Inference Attack | PerLTQA | ROC-AUC100 | 24 | |
| Agent Memory Question Answering | PerLTQA (test) | BLEU42.68 | 18 | |
| Memory Retrieval | PerLTQA CN | ERC93.12 | 14 | |
| Memory Retrieval | PerLTQA EN | ERC90.47 | 14 | |
| Long-term dialogue memory | PerLTQA (test) | Accuracy93.14 | 11 | |
| Long-horizon conversation utility evaluation | PerltQA | Accuracy80.62 | 6 | |
| Retrieval | PerLTQA | Ra@574.5 | 1 | |
| Proactive Assistant Evaluation | PerLTQA Category (test) | Response Frequency15 | 1 |
Showing 8 of 8 rows