We investigate whether pre-training exclusively on dialogue data results in formally and functionally apt small language models. Based on this pre-trained llamalogue model, we employ a variety of fine-tuning strategies to enforce "more communicative" text generations by our models. Although our models underperform on most standard BabyLM benchmarks, they excel at dialogue continuation prediction in a minimal pair setting. While PPO fine-tuning has mixed to adversarial effects on our models, DPO fine-tuning further improves their performance on our custom dialogue benchmark.
@article{arxiv.2510.20358,
title = {Dialogue Is Not Enough to Make a Communicative BabyLM (But Neither Is Developmentally Inspired Reinforcement Learning)},
author = {Francesca Padovani and Bastian Bunzeck and Manar Ali and Omar Momen and Arianna Bisazza and Hendrik Buschmeier and Sina Zarrieß},
journal= {arXiv preprint arXiv:2510.20358},
year = {2025}
}