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 aim to align large language models with the ever-changing,complex, and diverse human values (e.g., social norms) across time andlocations. This presents a challenge to existing alignment techniques, such assupervised fine-tuning, which internalize values within model parameters. Toovercome this, we propose an On-the-fly Preference Optimization (OPO) method,which is a real-time alignment that works in a streaming way. It employs anexternal memory to store established rules for alignment, which can constrainLLMs' behaviors without further training, allowing for convenient updates andcustomization of human values. We also introduce a scalable evaluation toassess the proposed method more effectively. Experimental results on bothhuman-annotated and auto-generated questions from legal and moral domainsindicate the effectiveness of the proposed OPO method. Our code and data arereleased at https://github.com/GAIR-NLP/OPO.

Align on the Fly: Adapting Chatbot Behavior to Established Norms