Regularizing Treatment Leakage for LLM-based Uplift Estimation

Bojing Feng, Jinqiu Deng, Zihan Hong, Yuan Xue, Wenqi Xiang, Shuhan Wang, Feng Xiao, Shiwei Wang, Zijiang Liu, Shaomin Li, Yongzhi Qi

Abstract

Large language models (LLMs) provide semantic features for uplift estimation, but treatment-revealing text can induce prediction shortcuts that do not transfer to deployment. We propose LLM-based Uplift Leakage Regularization (LULR) to address this risk. With the LLM encoder frozen, LULR penalizes dependence between the causal head's trainable shared representation and treatment while retaining separate outcome heads. The penalty uses a variational approximation to the contrastive log-ratio upper bound (CLUB) of mutual information, targeting treatment dependence rather than generic parameter shrinkage. Controlled proxy experiments provide exploratory evidence of improved ranking scores under injected treatment information relative to no regularization and the tested weight-decay setting. Structured-only benchmarks show modest, dataset-dependent effects. On an industrial pricing randomized controlled trial, mean offline [email protected] is 0.258 for LULR, compared with 0.106 for the same-backbone baseline. A separate one-week online A/B test of the complete deployed pricing policy records 5.45% higher sales volume and 1.91% higher gross merchandise value than the control arm. These short-term policy-level gains do not isolate the regularizer's effect.
BibTeX
@misc{feng2026regularizing,
  author = {Bojing Feng and Jinqiu Deng and Zihan Hong and Yuan Xue and Wenqi Xiang and Shuhan Wang and Feng Xiao and Shiwei Wang and Zijiang Liu and Shaomin Li and Yongzhi Qi},
  title = {{Regularizing Treatment Leakage for LLM-based Uplift Estimation}},
  year = {2026},
  eprint = {2611.20794},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2611.20794},
}
Figure