PAD: Personalized Alignment at Decoding-time. ICLR 2025

This repo contains the personalized reward model (PerRM) for alignment.

Our paper: https://openreview.net/pdf?id=e7AUJpP8bV

Github: https://github.com/zjuruizhechen/PAD

Downloads last month
1
Safetensors
Model size
8B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for RuizheChen/PAD

Finetuned
(515)
this model