4 papers
Looking in the Mirror: Introspecting Side-Effect Misalignments Induced by Fine-Tuning
Kotaro Yoshida, Laura Gomezjurado Gonzalez, Yukinori Yamamoto +3
Fine-tuning enables a source model to acquire desired capabilities and behaviors in a target domain while retaining much of its general-purpose competence. However, this adaptation…
The Long Delay to Arithmetic Generalization: When Learned Representations Outrun Behavior
Laura Gomezjurado Gonzalez
Grokking in transformers trained on algorithmic tasks is characterized by a long delay between training-set fit and abrupt generalization, but the source of that delay remains poor…
Orth-Dion: Eliminating Geometric Mismatch in Distributed Low-Rank Spectral Optimization
Tatsuhiro Nakamori, Laura Gomezjurado Gonzalez, Ganesh Talluri +5
Low-rank gradient compression reduces communication in distributed training by representing updates with rank- factors. Dion is a recent method that approximates Muon, a spectra…
On Fairness of Task Arithmetic: The Role of Task Vectors
Hiroki Naganuma, Kotaro Yoshida, Laura Gomezjurado Gonzalez +3
Model editing techniques, particularly task arithmetic with task vectors, offer an efficient alternative to full fine-tuning by enabling direct parameter updates through simple ari…