Cross-distribution performance on SAIR official benchmark (hard3)
81.3Accuracyhard3_overfitted
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| hard3_overfittedParticipant=Woon Siang Yi2026.04 | 81.3 | 83.1 | 30.3 | |
| Test 8Participant=Debtirtha Saha2026.04 | 71 | 76 | 11 | |
| bank_lookup_v8Participant=Arjun Garg2026.04 | 70.8 | 74.6 | 12.2 | |
| chat_v32Participant=Devanshu Dixit2026.04 | 68 | 74.3 | 18.5 | |
| AN38Participant=Cazares2026.04 | 65.3 | 67.6 | 24.3 | |
| traditional_mlParticipant=Reza Jamei2026.04 | 61.3 | 68.6 | 27.7 | |
| bank_lookup_v5Participant=Garg2026.04 | 60.8 | 69.3 | 9.3 | |
| distilled-r-12Participant=Heath2026.04 | 59 | 62.6 | 21 | |
| combo_jackParticipant=Reza Jamei2026.04 | 58.5 | 66.5 | 27.5 | |
| Stage 1 PromptParticipant=SimonRJ2026.04 | 58.3 | 67.3 | 0.3 | |
| 98_hard200Participant=Betka2026.04 | 56.3 | 55.9 | 42.7 | |
| hard98_derivate_2Participant=Pandey2026.04 | 55.5 | 53.9 | 41.5 | |
| AN45cParticipant=Cazares2026.04 | 55.5 | 63.7 | 7.5 |