Augmenting interpretable models with large language models during training

Method and contribution

Aug-Tree uses language-model-generated expansions to construct interpretable text features for tree splits.

Contrasts training-time feature expansion with Tree Prompting’s language-model calls while traversing a tree.

Read in context

LLM decision trees: from Tree Prompting to agentic classification →

Citation

Chandan Singh, Armin Askari, Rich Caruana, Jianfeng Gao. Augmenting interpretable models with large language models during training. Nature Communications 14, 7913.

@article{augtree,
  title = {{Augmenting interpretable models with large language models during training}},
  author = {Singh, Chandan and Askari, Armin and Caruana, Rich and Gao, Jianfeng},
  year = {2023},
  journal = {Nature Communications},
  url = {https://www.nature.com/articles/s41467-023-43713-1},
  volume = {14},
  number = {1},
  pages = {7913},
  doi = {10.1038/s41467-023-43713-1}
}