DogeRM: Equipping Reward Models with Domain Knowledge through Model Merging Paper • 2407.01470 • Published Jul 1 • 5