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 work we explore recent advances in instruction-tuning language modelson a range of open instruction-following datasets. Despite recent claims thatopen models can be on par with state-of-the-art proprietary models, theseclaims are often accompanied by limited evaluation, making it difficult tocompare models across the board and determine the utility of various resources.We provide a large set of instruction-tuned models from 6.7B to 65B parametersin size, trained on 12 instruction datasets ranging from manually curated(e.g., OpenAssistant) to synthetic and distilled (e.g., Alpaca) andsystematically evaluate them on their factual knowledge, reasoning,multilinguality, coding, and open-ended instruction following abilities througha collection of automatic, model-based, and human-based metrics. We furtherintroduce T\"ulu, our best performing instruction-tuned model suite finetunedon a combination of high-quality open resources. Our experiments show thatdifferent instruction-tuning datasets can uncover or enhance specific skills,while no single dataset (or combination) provides the best performance acrossall evaluations. Interestingly, we find that model and human preference-basedevaluations fail to reflect differences in model capabilities exposed bybenchmark-based evaluations, suggesting the need for the type of systemicevaluation performed in this work. Our evaluations show that the best model inany given evaluation reaches on average 87% of ChatGPT performance, and 73% ofGPT-4 performance, suggesting that further investment in building better basemodels and instruction-tuning data is required to close the gap. We release ourinstruction-tuned models, including a fully finetuned 65B T\"ulu, along withour code, data, and evaluation framework athttps://github.com/allenai/open-instruct to facilitate future research.

How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources