Metraipolis › archive › dc_779613f4ffd3
Building Telecom Language Models with RSOD: Domain Data, Specialist Post-Training, Memory-Conscious Distillation, and Networking Evaluation — arXiv:2610.27993
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.27993
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-21T07:06:58.408Z
- SHA-256 of the archived text
cad66059d96b07cdbbe26971727f524036e2f636470603cf117f62f032365069- Stable URL of this page
https://metraipolis.com/documento/dc_779613f4ffd3
How to cite
Building Telecom Language Models with RSOD: Domain Data, Specialist Post-Training, Memory-Conscious Distillation, and Networking Evaluation — arXiv:2610.27993 — arXiv (en), https://arxiv.org/abs/2610.27993 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_779613f4ffd3 (Metraipolis archive dc_779613f4ffd3, sha256 cad66059d96b07cdbbe26971727f524036e2f636470603cf117f62f032365069)
BibTeX
@misc{dc779613f4ffd3,
title = {Building Telecom Language Models with RSOD: Domain Data, Specialist Post-Training, Memory-Conscious Distillation, and Networking Evaluation — arXiv:2610.27993},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.27993},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_779613f4ffd3; sha256 cad66059d96b07cdbbe26971727f524036e2f636470603cf117f62f032365069},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27993. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Telecom language models must combine broad networking knowledge with specialist behavior across routing, broadband services, access technologies, and operational diagnostics while remaining practical to deploy as a single model. We present an end-to-end framework that builds source-grounded synthetic training data from structured networking taxonomies and expert graphs, materializes 212,559 routed training records across fourteen networking routes, trains six frozen specialist LoRA teachers over one shared continued-pretrained backbone, and consolidates their behavior into one deployable student adapter using Routed Specialist On-Policy Distillation (RSOD). In RSOD, the committed student generates each trajectory, the route-assigned specialist teacher-forces the exact native token sequence, compact teacher hidden states are reconstructed through the shared frozen output head, and only the student LoRA is updated. A controlled reconstruction validation shows mean absolute logit error of 7.26 x 10^-6 with 100% top-1 agreement between native and reconstructed teacher logits. On four RTX 6000 Ada GPUs, the completed 2,000-step run produced 16,000 trajectories with zero observed student-version or token- ID mismatches. Under the primary route-conditioned evaluation over 2,850 held-out questions, the raw Meta-Llama-3.1-8B-Instruct model scores 12.8661/25 micro (13.0172/25 macro), the combined routed six-teacher system scores 21.7641/25 micro, and the consolidated RSOD student, released as Llama-CalixCLM v1, reaches 21.7112/25 micro (21.8006/25 macro) on the same five dimension LLM-judge rubric. Under a separate neutral deployment-like prompt, the same student reaches 20.0028/25 micro (20.0778/25 macro). The consolidated student remains within 0.0529 micro-score points of the combined routed-teacher system while retaining a numerically validated and auditable train-serve execution path. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.27993 — 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 2026-10-21T07:05:43.461Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).