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

Retrieval-augmented generation (RAG) has become a main technique foralleviating hallucinations in large language models (LLMs). Despite theintegration of RAG, LLMs may still present unsupported or contradictory claimsto the retrieved contents. In order to develop effective hallucinationprevention strategies under RAG, it is important to create benchmark datasetsthat can measure the extent of hallucination. This paper presents RAGTruth, acorpus tailored for analyzing word-level hallucinations in various domains andtasks within the standard RAG frameworks for LLM applications. RAGTruthcomprises nearly 18,000 naturally generated responses from diverse LLMs usingRAG. These responses have undergone meticulous manual annotations at both theindividual cases and word levels, incorporating evaluations of hallucinationintensity. We not only benchmark hallucination frequencies across differentLLMs, but also critically assess the effectiveness of several existinghallucination detection methodologies. Furthermore, we show that using ahigh-quality dataset such as RAGTruth, it is possible to finetune a relativelysmall LLM and achieve a competitive level of performance in hallucinationdetection when compared to the existing prompt-based approaches usingstate-of-the-art large language models such as GPT-4.

RAGTruth: A Hallucination Corpus for Developing Trustworthy Retrieval-Augmented Language Models