Aggregate Benchmark Evaluation on FineTasks Arabic (test)
0.161ScoreAraMix-Matched
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AraMix-MatchedAblation=none2025.12 | 0.161 | 4.5 | |
| ArabicWeb24baseline=true2025.12 | 0.154 | — | |
| AraMix-MinHashAblation=none2025.12 | 0.153 | -0.6 | |
| AraMix-MatchedAblation=exclude content unique to baseline source2025.12 | 0.152 | -1.3 | |
| AraMix-MinHashAblation=exclude content unique to baseline source2025.12 | 0.144 | -6.5 |