Step-Back Profiling: Distilling User History for Personalized Scientific Writing

TL;DR

This paper introduces STEP-BACK PROFILING, a method that personalizes large language models for scientific writing by distilling user history into concise profiles, demonstrated on a new dataset with improved performance over baselines.

Contribution

The paper presents a novel profiling approach for LLM personalization in scientific writing and introduces the PSW dataset for multi-user experiments.

Findings

Effective user profiling improves personalized scientific writing.

Our method outperforms baselines by up to 3.6 points on LaMP.

Ablation studies confirm the importance of each component.

Abstract

Large language models (LLM) excel at a variety of natural language processing tasks, yet they struggle to generate personalized content for individuals, particularly in real-world scenarios like scientific writing. Addressing this challenge, we introduce STEP-BACK PROFILING to personalize LLMs by distilling user history into concise profiles, including essential traits and preferences of users. To conduct the experiments, we construct a Personalized Scientific Writing (PSW) dataset to study multi-user personalization. PSW requires the models to write scientific papers given specialized author groups with diverse academic backgrounds. As for the results, we demonstrate the effectiveness of capturing user characteristics via STEP-BACK PROFILING for collaborative writing. Moreover, our approach outperforms the baselines by up to 3.6 points on the general personalization benchmark (LaMP), including 7 personalization LLM tasks. Our ablation studies validate the contributions of different components in our method and provide insights into our task definition. Our dataset and code are available at \url{https://github.com/gersteinlab/step-back-profiling}.