List your company

Open-ended generation (Medical Diagnostic QA) on RJUA (test)

33.7Rouge-LW-REINFORCE
22.7825.61528.4531.285May 27, 2026

Evaluation results

MethodLinks
W-REINFORCEw+=0.2, w-=12026.0533.798.337.299.767.2
W-REINFORCEw+=1, w-=52026.0533.597.637.199.566.9
EAPO2026.053397.33699.566.4
NSR2026.0532.797.136.19966.2
W-REINFORCEw+=0.1, w-=12026.0532.295.836.699.866.1
GRPO2026.0531.297.734.498.765.5
W-REINFORCEw+=5, w-=12026.0530.29733.598.964.9
W-REINFORCEw+=1, w-=12026.0530.19732.699.464.8
W-REINFORCEw+=1, w-=0.052026.0529.993.332.498.363.5
PSR2026.0529.39432.797.963.5
W-REINFORCEw+=1, w-=0.22026.0527.493.128.897.861.8
W-REINFORCEw+=0.05, w-=12026.0526.390.334.699.662.7
W-REINFORCEw+=1, w-=0.12026.0523.292.226.497.359.8