You've Changed: Detecting Modification of Black-Box Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dima, Alden, Foulds, James, Pan, Shimei, Feldman, Philip
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908323281895424
author Dima, Alden
Foulds, James
Pan, Shimei
Feldman, Philip
author_facet Dima, Alden
Foulds, James
Pan, Shimei
Feldman, Philip
contents Large Language Models (LLMs) are often provided as a service via an API, making it challenging for developers to detect changes in their behavior. We present an approach to monitor LLMs for changes by comparing the distributions of linguistic and psycholinguistic features of generated text. Our method uses a statistical test to determine whether the distributions of features from two samples of text are equivalent, allowing developers to identify when an LLM has changed. We demonstrate the effectiveness of our approach using five OpenAI completion models and Meta's Llama 3 70B chat model. Our results show that simple text features coupled with a statistical test can distinguish between language models. We also explore the use of our approach to detect prompt injection attacks. Our work enables frequent LLM change monitoring and avoids computationally expensive benchmark evaluations.
format Preprint
id arxiv_https___arxiv_org_abs_2504_12335
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle You've Changed: Detecting Modification of Black-Box Large Language Models
Dima, Alden
Foulds, James
Pan, Shimei
Feldman, Philip
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) are often provided as a service via an API, making it challenging for developers to detect changes in their behavior. We present an approach to monitor LLMs for changes by comparing the distributions of linguistic and psycholinguistic features of generated text. Our method uses a statistical test to determine whether the distributions of features from two samples of text are equivalent, allowing developers to identify when an LLM has changed. We demonstrate the effectiveness of our approach using five OpenAI completion models and Meta's Llama 3 70B chat model. Our results show that simple text features coupled with a statistical test can distinguish between language models. We also explore the use of our approach to detect prompt injection attacks. Our work enables frequent LLM change monitoring and avoids computationally expensive benchmark evaluations.
title You've Changed: Detecting Modification of Black-Box Large Language Models
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2504.12335