Improving Diversity in Language Models: When Temperature Fails, Change the Loss

Abstract

Increasing diversity in language models is a challenging yet essential objective. A common approach is to raise the decoding temperature. In this work, we investigate this approach through a simplistic yet common case to provide insights into why decreasing temperature can improve quality (Precision), while increasing it often fails to boost coverage (Recall). Our analysis reveals that for a model to be effectively tunable through temperature adjustments, it must be trained toward coverage. To address this, we propose rethinking loss functions in language models by leveraging the Precision-Recall framework. Our results demonstrate that this approach achieves a substantially better trade-off between Precision and Recall than merely combining negative log-likelihood training with temperature scaling. These findings offer a pathway toward more versatile and robust language modeling techniques.

Cite

Text

Verine et al. "Improving Diversity in Language Models: When Temperature Fails, Change the Loss." Proceedings of the 42nd International Conference on Machine Learning, 2025.

Markdown

[Verine et al. "Improving Diversity in Language Models: When Temperature Fails, Change the Loss." Proceedings of the 42nd International Conference on Machine Learning, 2025.](https://mlanthology.org/icml/2025/verine2025icml-improving/)

BibTeX

@inproceedings{verine2025icml-improving,
  title     = {{Improving Diversity in Language Models: When Temperature Fails, Change the Loss}},
  author    = {Verine, Alexandre and Le Bronnec, Florian and Zheng, Kunhao and Allauzen, Alexandre and Chevaleyre, Yann and Negrevergne, Benjamin},
  booktitle = {Proceedings of the 42nd International Conference on Machine Learning},
  year      = {2025},
  pages     = {61266-61300},
  volume    = {267},
  url       = {https://mlanthology.org/icml/2025/verine2025icml-improving/}
}