Metraipolis › archive › dc_286fe1fb5ada
Saidzoda-Tajik-LFM-2.6B: Adapting a 2.6B Hybrid Model to Tajik Tool Calling, and What Continued Pretraining Takes Away — arXiv:2610.41576
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.41576
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T07:54:42.571Z
- SHA-256 of the archived text
a971369f5c07b8b451e3e9964d7438de1d70cc1af94c9d26409065adfbe2c1ea- Stable URL of this page
https://metraipolis.com/documento/dc_286fe1fb5ada
How to cite
Saidzoda-Tajik-LFM-2.6B: Adapting a 2.6B Hybrid Model to Tajik Tool Calling, and What Continued Pretraining Takes Away — arXiv:2610.41576 — arXiv (en), https://arxiv.org/abs/2610.41576 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_286fe1fb5ada (Metraipolis archive dc_286fe1fb5ada, sha256 a971369f5c07b8b451e3e9964d7438de1d70cc1af94c9d26409065adfbe2c1ea)
BibTeX
@misc{dc286fe1fb5ada,
title = {Saidzoda-Tajik-LFM-2.6B: Adapting a 2.6B Hybrid Model to Tajik Tool Calling, and What Continued Pretraining Takes Away — arXiv:2610.41576},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.41576},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_286fe1fb5ada; sha256 a971369f5c07b8b451e3e9964d7438de1d70cc1af94c9d26409065adfbe2c1ea},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41576. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We release Saidzoda-Tajik-LFM-2.6B, a Tajik adaptation of the 2.6B-parameter hybrid model LFM2.5-2.6B-Base, built for tool calling in Tajik on modest hardware. The recipe has two stages: LoRA continued pretraining (CPT) on 1,485,319 cleaned Tajik web documents for 3 epochs, then one epoch of supervised fine-tuning (SFT) on a mix of machine-translated Tajik tool calls, English tool calls kept as replay, and Tajik instructions. On a 174-item tool-calling test set built from BFCL v3 and machine-translated to Tajik, strict full-call accuracy on Tajik prompts rises from 20.69 for the base model to 64.14 (paired difference +43.45 points, 95% interval +35.2 to +51.7); a lenient parser that also accepts the base model's own call formats puts the base at 31.72. English shows no detectable change (76.55 and 77.24). Every CPT checkpoint scores zero: CPT removes the native tool-call format and, as the lenient re-scoring shows, part of the skill itself. Outside tool calling, accuracy on a Tajik knowledge benchmark rises from 28.0 to 35.0, Belebele Tajik shows no significant change, and English MMLU falls from 60.8 to 52.5. The model is not a general Tajik assistant. It almost never declines when no tool fits (6.9% correct on Tajik, against 31.0% for the base under the lenient parser), it is weak at multi-turn calls, and its calls should be checked before they are executed. We describe the data, the training settings and every evaluation in enough detail to repeat them. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.41576 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T07:49:46.003Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).