Question Answering on XQuAD Code Switched P, I - (EN), Q(X) (Avg)
100Language ConsistencyQwen-3-32B (Full scope SFT)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen-3-32B (Full scope SFT)Model=Qwen-3-32B, Fine-tuning Strategy=Full scope SFT2026.01 | 100 | 51.87 | |
| Qwen-3-32B (Selective SFT)Model=Qwen-3-32B, Fine-tuning Strategy=Selective SFT2026.01 | 100 | 53.53 | |
| Bloom-7.1B (Full scope SFT)Model=Bloom-7.1B, Fine-tuning Strategy=Full scope SFT2026.01 | 99.89 | 21.03 | |
| Bloom-7.1B (Selective SFT)Model=Bloom-7.1B, Fine-tuning Strategy=Selective SFT2026.01 | 99.8 | 21.02 | |
| Qwen-3-32B (Random Selective SFT)Model=Qwen-3-32B, Fine-tuning Strategy=Random Selective SFT2026.01 | 97.93 | 1.1 | |
| Bloom-7.1B (Pre-Finetuning)Model=Bloom-7.1B, Fine-tuning Strategy=Pre-Finetuning2026.01 | 71.23 | 6.58 | |
| Bloom-7.1B (Random Selective SFT)Model=Bloom-7.1B, Fine-tuning Strategy=Random Selective SFT2026.01 | 43.11 | 0 | |
| Qwen-3-32B (Pre-Finetuning)Model=Qwen-3-32B, Fine-tuning Strategy=Pre-Finetuning2026.01 | 11.01 | 52.65 |