Signal2026-06-09
Papers With Code

Self-Evaluation Is Already There: Eliciting Latent Judge Calibration in Base LLMs with Minimal Data

Open primary sourceRead on Substack