Chemicals and Drugs Classification on PubMed synthetic data
75.48AccuracyFederated Controllable Text Generation with Refinement
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Federated Controllable Text Generation with RefinementBias=D, C, M, N, Privacy budget (epsilon)=∞, Refinement=true2026.03 | 75.48 | 75.44 | |
| Federated Controllable Text Generation with RefinementBias=C, M, N, Privacy budget (epsilon)=∞, Refinement=true2026.03 | 73.04 | 72.54 | |
| Federated Controllable Text Generation with RefinementBias=D, C, M, N, Privacy budget (epsilon)=8, Refinement=true2026.03 | 70.47 | 68.28 | |
| Federated Controllable Text Generation with RefinementBias=D, C, M, N, Privacy budget (epsilon)=∞, Refinement=false2026.03 | 69.8 | 69.84 | |
| Federated Controllable Text Generation with RefinementBias=M, N, Privacy budget (epsilon)=∞, Refinement=true2026.03 | 68.4 | 68.19 | |
| Federated Controllable Text Generation with RefinementBias=C, M, N, Privacy budget (epsilon)=8, Refinement=true2026.03 | 66.38 | 64.29 | |
| Federated Controllable Text Generation with RefinementBias=D, C, M, N, Privacy budget (epsilon)=8, Refinement=false2026.03 | 63.38 | 61.35 | |
| Federated Controllable Text Generation with RefinementBias=C, M, N, Privacy budget (epsilon)=∞, Refinement=false2026.03 | 60.52 | 57.7 | |
| Federated Controllable Text Generation with RefinementBias=M, N, Privacy budget (epsilon)=∞, Refinement=false2026.03 | 56.96 | 52.17 | |
| Federated Controllable Text Generation with RefinementBias=M, N, Privacy budget (epsilon)=8, Refinement=true2026.03 | 56.62 | 53.37 | |
| Federated Controllable Text Generation with RefinementBias=C, M, N, Privacy budget (epsilon)=8, Refinement=false2026.03 | 54.37 | 52.64 | |
| Federated Controllable Text Generation with RefinementBias=M, N, Privacy budget (epsilon)=8, Refinement=false2026.03 | 50.34 | 46.88 |