Live page · Day archive

Judging in Latent Space: Efficient Generative Reward Modeling via Semantics-Preserving Compression

Read the original at HF Daily Papers

Summary

Researchers introduce LatentGRM, a framework that compresses evaluation trajectories by 8.9 to 9.2 times to reduce inference time while maintaining preference accuracy.

Carried by: HF Daily Papers. First seen: 2026-10-12T02:30:09Z.