A Causal Language Modeling Detour Improves Encoder Continued Pretraining

Rian Touchent, Éric de la ClergeriearXiv 2026 · DOI

BibTeX
@misc{touchent2026causal,
  author = {Rian Touchent and Éric de la Clergerie},
  title = {{A Causal Language Modeling Detour Improves Encoder Continued Pretraining}},
  year = {2026},
  eprint = {2605.12438},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2605.12438},
}

Loading related content...

Figure