Doppelg\"anger's Watch: A Split Objective Approach to Large Language Models
Computation and Language
2024-09-11 v1 Artificial Intelligence
Abstract
In this paper, we investigate the problem of "generation supervision" in large language models, and present a novel bicameral architecture to separate supervision signals from their core capability, helpfulness. Doppelg\"anger, a new module parallel to the underlying language model, supervises the generation of each token, and learns to concurrently predict the supervision score(s) of the sequences up to and including each token. In this work, we present the theoretical findings, and leave the report on experimental results to a forthcoming publication.
Cite
@article{arxiv.2409.06107,
title = {Doppelg\"anger's Watch: A Split Objective Approach to Large Language Models},
author = {Shervin Ghasemlou and Ashish Katiyar and Aparajita Saraf and Seungwhan Moon and Mangesh Pujari and Pinar Donmez and Babak Damavandi and Anuj Kumar},
journal= {arXiv preprint arXiv:2409.06107},
year = {2024}
}