Future of Personalized, Aligned Language Models
Name
han-swhan-sm-eecs-2025-thesis.pdf
Description
Thesis PDF
Size
2.47 MB
Format
Adobe PDF
Checksum (MD5)
5e59dbabd267507da3bdfb1db3fc0b46
Author(s)
Han, Seungwook
Advisor(s)
Agrawal, Pulkit
Date Issued
May 2025
Publisher
Massachusetts Institute of Technology
Abstract
Aligning Large Language Models (LLMs) to cater to different human preferences, learning new skills, and unlearning harmful behavior is an important problem. Search-based methods, such as Best-of-N or Monte-Carlo Tree Search, are effective, but impractical for LLM adaptation due to their high inference cost. On the other hand, using Reinforcement Learning (RL) for adaptation is computationally efficient, but performs worse due to the optimization challenges in co-training the value function and the policy. We present a new framework for reward optimization, Value Augmented Sampling (VAS), that can maximize different reward functions using data sampled from only the initial, frozen LLM. VAS solves for the optimal reward-maximizing policy without co-training the policy and the value function, making the optimization stable, outperforming established baselines, such as PPO and DPO, on standard benchmarks, and achieving comparable results to Best-of-128 with lower inference cost. Unlike existing RL methods that require changing the weights of the LLM, VAS does not require access to the weights of the pre-trained LLM. Thus, it can even adapt LLMs (e.g., ChatGPT), which are available only as APIs. In addition, our algorithm unlocks the new capability of composing several rewards and controlling the extent of each one during deployment time. By bringing together stability, flexibility, and efficiency, we explore the future of aligned, personalized language models that can be adapted seamlessly to meet a wide spectrum of human preferences.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
