Question Answering on LongMemEval 500 questions
61.4QA AccuracyDiagnostic-conditional
Evaluation Results
| Method | Links | |
|---|---|---|
| Diagnostic-conditionalScale=8B2026.05 | 61.4 | |
| Diagnostic-conditionalScale=14B2026.05 | 60 | |
| Best unconditionalScale=8B2026.05 | 59.8 | |
| Best unconditionalScale=14B2026.05 | 59.8 | |
| IntactScale=14B2026.05 | 58.6 | |
| Best unconditionalScale=4B2026.05 | 58.2 | |
| Diagnostic-conditionalScale=4B2026.05 | 57.4 | |
| IntactScale=8B2026.05 | 57.4 | |
| IntactScale=4B2026.05 | 56.2 | |
| Best unconditionalScale=0.6B2026.05 | 34.4 | |
| Diagnostic-conditionalScale=0.6B2026.05 | 34.4 | |
| IntactScale=0.6B2026.05 | 34 |