Plausible Token Amplification for Improving Accuracy of Differentially Private In-Context Learning Based on Implicit Bayesian Inference

Yusuke Yamasaki, Kenta Niwa, Daiki Chijiwa, Takumi Fukami, Takayuki Miura
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:70205-70248, 2025.

Abstract

We propose Plausible Token Amplification (PTA) to improve the accuracy of Differentially Private In-Context Learning (DP-ICL) using DP synthetic demonstrations. While Tang et al. empirically improved the accuracy of DP-ICL by limiting vocabulary space during DP synthetic demonstration generation, its theoretical basis remains unexplored. By interpreting ICL as implicit Bayesian inference on a concept underlying demonstrations, we not only provide theoretical evidence supporting Tang et al.’s empirical method but also introduce PTA, a refined method for modifying next-token probability distribution. Through the modification, PTA highlights tokens that distinctly represent the ground-truth concept underlying the original demonstrations. As a result, generated DP synthetic demonstrations guide the Large Language Model to successfully infer the ground-truth concept, which improves the accuracy of DP-ICL. Experimental evaluations on both synthetic and real-world text-classification datasets validated the effectiveness of PTA.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-yamasaki25a, title = {Plausible Token Amplification for Improving Accuracy of Differentially Private In-Context Learning Based on Implicit {B}ayesian Inference}, author = {Yamasaki, Yusuke and Niwa, Kenta and Chijiwa, Daiki and Fukami, Takumi and Miura, Takayuki}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {70205--70248}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/yamasaki25a/yamasaki25a.pdf}, url = {https://proceedings.mlr.press/v267/yamasaki25a.html}, abstract = {We propose Plausible Token Amplification (PTA) to improve the accuracy of Differentially Private In-Context Learning (DP-ICL) using DP synthetic demonstrations. While Tang et al. empirically improved the accuracy of DP-ICL by limiting vocabulary space during DP synthetic demonstration generation, its theoretical basis remains unexplored. By interpreting ICL as implicit Bayesian inference on a concept underlying demonstrations, we not only provide theoretical evidence supporting Tang et al.’s empirical method but also introduce PTA, a refined method for modifying next-token probability distribution. Through the modification, PTA highlights tokens that distinctly represent the ground-truth concept underlying the original demonstrations. As a result, generated DP synthetic demonstrations guide the Large Language Model to successfully infer the ground-truth concept, which improves the accuracy of DP-ICL. Experimental evaluations on both synthetic and real-world text-classification datasets validated the effectiveness of PTA.} }
Endnote
%0 Conference Paper %T Plausible Token Amplification for Improving Accuracy of Differentially Private In-Context Learning Based on Implicit Bayesian Inference %A Yusuke Yamasaki %A Kenta Niwa %A Daiki Chijiwa %A Takumi Fukami %A Takayuki Miura %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-yamasaki25a %I PMLR %P 70205--70248 %U https://proceedings.mlr.press/v267/yamasaki25a.html %V 267 %X We propose Plausible Token Amplification (PTA) to improve the accuracy of Differentially Private In-Context Learning (DP-ICL) using DP synthetic demonstrations. While Tang et al. empirically improved the accuracy of DP-ICL by limiting vocabulary space during DP synthetic demonstration generation, its theoretical basis remains unexplored. By interpreting ICL as implicit Bayesian inference on a concept underlying demonstrations, we not only provide theoretical evidence supporting Tang et al.’s empirical method but also introduce PTA, a refined method for modifying next-token probability distribution. Through the modification, PTA highlights tokens that distinctly represent the ground-truth concept underlying the original demonstrations. As a result, generated DP synthetic demonstrations guide the Large Language Model to successfully infer the ground-truth concept, which improves the accuracy of DP-ICL. Experimental evaluations on both synthetic and real-world text-classification datasets validated the effectiveness of PTA.
APA
Yamasaki, Y., Niwa, K., Chijiwa, D., Fukami, T. & Miura, T.. (2025). Plausible Token Amplification for Improving Accuracy of Differentially Private In-Context Learning Based on Implicit Bayesian Inference. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:70205-70248 Available from https://proceedings.mlr.press/v267/yamasaki25a.html.

Related Material