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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this paper, we tackle the emerging challenge of unintended harmful contentgeneration in Large Language Models (LLMs) with a novel dual-stage optimisationtechnique using adversarial fine-tuning. Our two-pronged approach employs anadversarial model, fine-tuned to generate potentially harmful prompts, and ajudge model, iteratively optimised to discern these prompts. In thisadversarial cycle, the two models seek to outperform each other in theprompting phase, generating a dataset of rich examples which are then used forfine-tuning. This iterative application of prompting and fine-tuning allowscontinuous refinement and improved performance. The performance of our approachis evaluated through classification accuracy on a dataset consisting ofproblematic prompts not detected by GPT-4, as well as a selection ofcontentious but unproblematic prompts. We show considerable increase inclassification accuracy of the judge model on this challenging dataset as itundergoes the optimisation process. Furthermore, we show that a rudimentarymodel \texttt{ada} can achieve 13\% higher accuracy on the hold-out test setthan GPT-4 after only a few rounds of this process, and that this fine-tuningimproves performance in parallel tasks such as toxic comment identification.

Adversarial Fine-Tuning of Language Models: An Iterative Optimisation Approach for the Generation and Detection of Problematic Content