Mathematical Reasoning on GSM (Pairwise Comparison)
48.9Win RateSeqXKD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SeqXKDEvaluation Model=GPT-4, Teacher Model=Llama-2-7b, Student Model=T5-large, Comparison Baseline=SeqKD2026.02 | 48.9 | 2.8 | 48.3 |
| Method | Links | |||
|---|---|---|---|---|
| SeqXKDEvaluation Model=GPT-4, Teacher Model=Llama-2-7b, Student Model=T5-large, Comparison Baseline=SeqKD2026.02 | 48.9 | 2.8 | 48.3 |