Can LLMs Predict Their Own Failures? Self-Awareness via Internal Circuits

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Ghasemabadi, Amirhosein, Niu, Di
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2512.20578