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/

Zendy Plus

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

Zendy Tools

Zendy Open

Real-life tasks such as giving legal or technical advice often lack completecontext at the outset and can have disparate answers depending thereon. Theability to derive missing factual information by asking clarifying questions(ACQ) is an important element of real-life collaboration on such reasoningtasks. Existing factual clarification question challenges evaluate generationsbased on word overlap or human evaluations. Recent work explores generating aresponse to the clarifying question then evaluating its utility directly. Sofar, these tasks are limited to disambiguating the user's intent rather thanconcrete facts about the situation. The factual domain presents uniquechallenges since responses to clarification questions must be factually truefor accurate evaluation. To enable evaluation of factual domain clarificationquestion generation, We present a new task that focuses on the ability toelicit missing information in multi-hop reasoning tasks. The task,HotpotQA-FLM, can be evaluated automatically, making it convenient forbenchmarking language models. We observe that humans outperform GPT-4 by alarge margin, while Llama 3 8B Instruct does not even beat the dummy baselinein some metrics. Finally, we find by fine-tuning Llama 3 8B Instruct on its owngenerations, filtered via rejection sampling, we can improve informationrecovery by 27.6 percent.

Alexpaca: Learning Factual Clarification Question Generation Without Examples