Email Writing on Email Writing (Strong User)
0.2406Mean Total Edit DistanceEarlyEnsemble
Evaluation Results
| Method | Links | |
|---|---|---|
| EarlyEnsembleTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.2406 | |
| DPOTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.2955 | |
| LateEnsembleTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.3123 | |
| DPOUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.3365 | |
| SFTTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.3377 | |
| EarlyEnsembleUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.3438 | |
| SFTUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.4159 | |
| LateEnsembleUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.4202 | |
| BaseTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.4241 | |
| BaseUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.5108 |