STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models
Jiliang Ni ∗, Jiachen Pu ∗, Zhongyi Yang ∗, Jingfeng Luo, Conggang Hu † · 2026
Read STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models by Jiliang Ni ∗, Jiachen Pu ∗, Zhongyi Yang ∗, Jingfeng Luo, Conggang Hu † (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-03022.txt.
Description
Jiliang Ni and 4 co-authors shrink language model skills into very small models in a 2026 paper. Their focus is function calling, where a model must invoke the right tools with the right arguments, a core skill for artificial intelligence agents. Large models do this well but are too big for widespread deployment on limited hardware.
Their framework, STAR, combines two innovations in one training curriculum. Constrained distillation suppresses confidently wrong predictions while preserving exploration. A similarity-guided reinforcement stage rewards outputs by how closely they match the correct call, giving a rich continuous signal. Tests on challenging benchmarks set new best results for the size classes. Their smallest model beat every open model under a billion parameters, and several larger ones.
Catalog record
| CREATOR | Jiliang Ni ∗, Jiachen Pu ∗, Zhongyi Yang ∗, Jingfeng Luo, Conggang Hu † |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-03022