sludge-informational-roberta

Fine-tuned RoBERTa-base classifier for detecting informational sludge (barriers to knowing — inadequate or obscured information) in consumer financial complaint narratives.

Developed and validated in (manuscript in preparation):

Chesterfield, A., Gillespie, A., Goddard, A. and Krpan, D. (2026). Feeling the Friction: Developing and validating text classifiers for sludge in consumer complaints.

Code and data: GitHub | Companion model: sludge-process-roberta

What this model detects

Informational sludge = barriers to knowing: inadequate, obscured, or confusing information provision, including organisational opacity where companies cannot or will not explain their decisions. Binary classification: 1 = informational sludge present, 0 = absent.

Training details

Parameter	Value
Base model	roberta-base
Training samples	825 (85% stratified split of n=971)
Test samples	146 (15% held-out, evaluated once)
Epochs	4
Learning rate	2e-5
Batch size	8
Warmup ratio	0.1
Class weighting	Balanced
Random seed	42

Performance (held-out test set, n=146)

Metric	Score
MCC	0.621
F1	0.667
Precision	0.778
Recall	0.583
Accuracy	0.904

How to use

from transformers import pipeline

classifier = pipeline("text-classification", model="AlexChesterfield/sludge-informational-roberta")
complaint = "No one could explain why my claim was denied or what I needed to do next."
result = classifier(complaint)
# LABEL_1 = sludge present, LABEL_0 = absent

Citation

@article{chesterfield2026sludge,
  title={Feeling the Friction: Developing and validating text classifiers for sludge in consumer complaints},
  author={Chesterfield, Alexandra and Gillespie, Alex and Goddard, Alex and Krpan, Dario},
  year={2026}
}