Fine-tuned from jordanpainter/diallm-gemma-cpt on standard ultrafeedback-binarized-preferences corpus, providing a general instruction-following objective with no explicit dialectal signal.

Code, checkpoints, preference datasets, linguistic-analysis toolkit: https://github.com/surrey-nlp/diallm

Paper: https://arxiv.org/abs/2607.07669

Citation

Base model: Gemma 3-4B-it
Target variety: broad — no variety targeting
Adaptation thread: implicit (broad)
Pipeline stage: SFT (from the CPT checkpoint)

bibtex

@article{painter2026diallm,
  title     = {DiaLLM: An Investigation into the Robustness-Generation Gap in English Dialect Adaptation},
  author    = {Painter, Jordan and Srirag, Dipankar and Kappiyath, Adarsh and Kanojia, Diptesh and Joshi, Aditya and Yin, Lu},
  year      = {2026},
  eprint    = {2607.07669},
  archivePrefix = {arXiv}
}

jordanpainter/diallm-gemma-sft-all

DiaLLM — Gemma 3-4B-it — Broad — SFT

Citation

jordanpainter/
diallm-gemma-sft-all