Learning to Zoom Efficiently with a Contrastive Curriculum

The paper introduces a contrastive curriculum for teaching large language models to use a zoom‑in tool without supervised fine‑tuning, leveraging an InfoNCE‑style intrinsic reward that progressively hardens negative examples. Experiments on V*, HRBench, and MME‑RealWorld demonstrate competitive performance and efficiency, with the method even surpassing baselines when replacing traditional SFT. A new synthetic dataset, Muffin&Chihuahua (M&C), is created to directly evaluate zoom‑in capability, revealing that recall best correlates with task success.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

• agents
• fine-tuning