lifan-yuan/PLMCalibration

Code for ACL 2023 paper "A Close Look into the Calibration of Pre-trained Language Models"

28
/ 100
Experimental

This project helps researchers and developers understand how reliable the predictions of Pre-trained Language Models (PLMs) are. It takes a PLM and a text classification dataset as input, then outputs detailed metrics showing how 'calibrated' the model's confidence scores are. Scientists and machine learning engineers working with language models for tasks like sentiment analysis or spam detection would use this.

No commits in the last 6 months.

Use this if you need to rigorously evaluate the confidence scores of your large language models, especially in research or when deploying models where trust in predictions is critical.

Not ideal if you are looking for a tool to directly improve or 'calibrate' your deployed language models without deep analysis of the underlying calibration dynamics.

natural-language-processing machine-learning-research model-evaluation text-classification language-model-deployment
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 16 / 25
Community 7 / 25

How are scores calculated?

Stars

11

Forks

1

Language

Python

License

MIT

Last pushed

May 09, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/lifan-yuan/PLMCalibration"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.