Neurobiber: Fast and Interpretable Stylistic Feature Extraction

TL;DR

Neurobiber is a transformer-based system that rapidly extracts interpretable stylistic features from texts, enabling large-scale stylistic analysis and authorship verification with high speed and competitive accuracy.

Contribution

It introduces Neurobiber, a fast, interpretable transformer-based tool for stylistic feature extraction based on Biber's MDA, with open-source implementation and broad applicability.

Findings

Up to 56 times faster than existing systems

Replicates classic MDA insights on the CORE corpus

Achieves competitive performance on PAN 2020 authorship verification

Abstract

Linguistic style is pivotal for understanding how texts convey meaning and fulfill communicative purposes, yet extracting detailed stylistic features at scale remains challenging. We present Neurobiber, a transformer-based system for fast, interpretable style profiling built on Biber's Multidimensional Analysis (MDA). Neurobiber predicts 96 Biber-style features from our open-source BiberPlus library (a Python toolkit that computes stylistic features and provides integrated analytics, e.g., PCA and factor analysis). Despite being up to 56 times faster than existing open source systems, Neurobiber replicates classic MDA insights on the CORE corpus and achieves competitive performance on the PAN 2020 authorship verification task without extensive retraining. Its efficient and interpretable representations readily integrate into downstream NLP pipelines, facilitating large-scale stylometric research, forensic analysis, and real-time text monitoring. All components are made publicly available.