English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Large language models (LLMs) have demonstrated powerful capabilities in bothtext understanding and generation. Companies have begun to offer Embedding as aService (EaaS) based on these LLMs, which can benefit various natural languageprocessing (NLP) tasks for customers. However, previous studies have shown thatEaaS is vulnerable to model extraction attacks, which can cause significantlosses for the owners of LLMs, as training these models is extremely expensive.To protect the copyright of LLMs for EaaS, we propose an Embedding Watermarkmethod called EmbMarker that implants backdoors on embeddings. Our methodselects a group of moderate-frequency words from a general text corpus to forma trigger set, then selects a target embedding as the watermark, and inserts itinto the embeddings of texts containing trigger words as the backdoor. Theweight of insertion is proportional to the number of trigger words included inthe text. This allows the watermark backdoor to be effectively transferred toEaaS-stealer's model for copyright verification while minimizing the adverseimpact on the original embeddings' utility. Our extensive experiments onvarious datasets show that our method can effectively protect the copyright ofEaaS models without compromising service quality.

Are You Copying My Model? Protecting the Copyright of Large Language Models for EaaS via Backdoor Watermark