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

Large Language Models (LLMs) are smart but forgetful. Recent studies, (e.g.,(Bubeck et al., 2023)) on modern LLMs have shown that they are capable ofperforming amazing tasks typically necessitating human-level intelligence.However, unlike humans, frozen LLMs do not improve over time; they neitheracquire new knowledge nor learn from their successes or failures. Someapproaches to improving the intelligence of LLMs include fine-tuning modelsbased on problem-solving performance (Zelikman et al., 2022), and buildingbigger and more sophisticated models (Bubeck et al., 2023). However, thesemethods have the drawback of requiring substantial data and computationalresources to retrain existing models. In this paper, we explore the use ofRetrieval Augmented Generation, also known as RAG (Lewis et al., 2021) toimprove problem-solving performance. We propose ARM-RAG (Auxiliary RationaleMemory for Retrieval Augmented Generation), a system that learns from itssuccesses without incurring high training costs. We demonstrate that thestorage and subsequent retrieval of reasoning chains have a positive influenceon performance in grade-school math problems.

Enhancing LLM Intelligence with ARM-RAG: Auxiliary Rationale Memory for Retrieval Augmented Generation