Sondej, F., & Yang, Y. (2025). Collapse of Irrelevant Representations (CIR) Ensures Robust and Non-Disruptive LLM Unlearning.
Cita Chicago Style (17a ed.)Sondej, Filip, y Yushi Yang. Collapse of Irrelevant Representations (CIR) Ensures Robust and Non-Disruptive LLM Unlearning. 2025.
Cita MLA (9a ed.)Sondej, Filip, y Yushi Yang. Collapse of Irrelevant Representations (CIR) Ensures Robust and Non-Disruptive LLM Unlearning. 2025.
Precaución: Estas citas no son 100% exactas.