Short-continuation text generation on Multi-domain short-continuation
0.402MMD^2MoEMA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MoEMAEvaluation feature space=GPT-2, Evaluation protocol=Averages over five random repeats, Weight type=input-adaptive2026.07 | 0.402 | 0.074 | 47.004 | |
| StaticMAEvaluation feature space=GPT-2, Evaluation protocol=Averages over five random repeats, Weight type=fixed-weight2026.07 | 0.412 | 0.077 | 47.809 | |
| SimpleAvgEvaluation feature space=GPT-2, Evaluation protocol=Averages over five random repeats2026.07 | 0.413 | 0.077 | 47.681 | |
| BestSingleEvaluation feature space=GPT-2, Evaluation protocol=Averages over five random repeats2026.07 | 0.432 | 0.082 | 77.606 |