Children’s English Reading Story Generation via Supervised Fine-Tuning of Compact LLMs with Controllable Difficulty and Safety
Qian ShenUniversity of Florida [email protected], Fanghua CaoUniversity of Florida [email protected], Min YaoUniversity of Florida [email protected] Shlok GildaUniversity of Florida [email protected] Bonnie J. DorrUniversity of Florida [email protected] Walter L. LeiteUniversity of Florida [email protected] · 2026
Read Children’s English Reading Story Generation via Supervised Fine-Tuning of Compact LLMs with Controllable Difficulty and Safety by Qian ShenUniversity of Florida [email protected], Fanghua CaoUniversity of Florida [email protected], Min YaoUniversity of Florida [email protected] Shlok GildaUniversity of Florida [email protected] Bonnie J. DorrUniversity of Florida [email protected] Walter L. LeiteUniversity of Florida [email protected] (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-2605-13709.txt.
Description
Qian Shen and five colleagues fine-tune three 8B-parameter language models for children's English reading story generation with controllable difficulty and safety. Using an expert-designed reading curriculum plus stories from GPT-4o and Llama 3.3 70B, the team tests fine-tuning designs for compact affordable models. With apt fine-tuning, 8B models beat zero-shot GPT-4o and Llama 3.3 70B on difficulty metrics, with almost no visible safety issues. Such fine-tuned models could serve teachers, parents and children at school and home.
Catalog record
| CREATOR | Qian ShenUniversity of Florida [email protected], Fanghua CaoUniversity of Florida [email protected], Min YaoUniversity of Florida [email protected] Shlok GildaUniversity of Florida [email protected] Bonnie J. DorrUniversity of Florida [email protected] Walter L. LeiteUniversity of Florida [email protected] |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44736 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.13709