Summarization on Summarization (Weak User)
0.2577Mean Total Edit DistanceEarlyEnsemble
Evaluation Results
| Method | Links | |
|---|---|---|
| EarlyEnsembleTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.2577 | |
| LateEnsembleTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.3403 | |
| EarlyEnsembleUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.3586 | |
| LateEnsembleUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.4403 | |
| SFTTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.7698 | |
| DPOTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.78 | |
| DPOUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.8267 | |
| BaseTest-time User Model=Llama-3.3-70B-Instruct, Online learning phase=true2026.01 | 0.8558 | |
| SFTUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.9304 | |
| BaseUser LLM=Qwen-3 32B, Seeds=3, Learning Phase=Online2026.01 | 0.9445 |