ControlLM: Crafting Diverse Personalities for Language Models

TL;DR

ControlLM enables real-time, inference-time control of language model personalities, allowing for diverse, human-like behaviors and improved task performance without additional training.

Contribution

This work introduces ControlLM, a novel method leveraging differential activation patterns to control language model personalities at inference time.

Findings

ControlLM can elicit diverse persona behaviors without training.

It allows precise personality control matching human values.

Enhanced reasoning and question answering through attribute amplification.

Abstract

As language models continue to scale in size and capability, they display an array of emerging behaviors, both beneficial and concerning. This heightens the need to control model behaviors. We hope to be able to control the personality traits of language models at the inference-time so as to have various character features, on top of which the requirements of different types of tasks can be met. Personality is a higher-level and more abstract behavioral representation for language models. We introduce ControlLM, which leverages differential activation patterns, derived from contrasting behavioral prompts in the model's latent space, to influence the model's personality traits at inference. This approach allows for the precise, real-time adjustment of model behavior. First, we demonstrate ControlLM's capacity to elicit diverse persona behaviors without any training, while precision control allows personality traits to closely match average human values. Subsequently, we showcase improved reasoning and question answering through selective amplification of beneficial attributes like conscientiousness and friendliness. We hope that this work will inspire research on controlling human-like behaviors of language models and provide insights for future research. Our code is publicly available at: https://github.com/wengsyx/ControlLM.