Purging the Gray Zone: Latent-Geometric Denoising for Precise Knowledge Boundary Awareness

ArXi:2604.14324v1 Announce Type: new Large language models (LLMs) often exhibit hallucinations due to their inability to accurately perceive their own knowledge boundaries. Existing abstention fine-tuning methods typically partition datasets directly based on response accuracy, causing models to suffer from severe label noise near the decision boundaries and consequently exhibit high rates of abstentions or hallucinations.