Preference evaluation on LLM-as-a-Judge comparison set
33.7TCRM Better RateTCRM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TCRMtraining_setup=PPO, tuning_mode=LoRA, comparison_partner=Baseline2026.04 | 33.7 | 9.2 | 34.3 | 22.8 |
| Method | Links | ||||
|---|---|---|---|---|---|
| TCRMtraining_setup=PPO, tuning_mode=LoRA, comparison_partner=Baseline2026.04 | 33.7 | 9.2 | 34.3 | 22.8 |