Skip to content
New · Literature Matrix

See how the research fits together

Search a topic, a paper title or a DOI. The Literature Matrix finds up to 40 real papers on OpenAlex, reads what each one is about with SPECTER2 and draws them as one map — similar papers side by side, topic groups in colour, the most-cited papers biggest.

Included in the Pro plan

Try a real map:
app/ai-tools/literature-matrix?q=Large+language+models+in+education

Literature Matrix

Large language models in education
Build map

18 of 20 new maps left today · maps someone already built open free

  • Artificial Intelligence in Education28
  • Natural Language Processing Techniques11

Topic groups

  • Artificial Intelligence in Education28
  • Natural Language Processing Techniques11

Most-cited papers

Click any circle for its details.

These are real maps the Literature Matrix built from OpenAlex — hover a paper to see its links, click it to open its details.

  • Up to 40 papers per map
  • 4 closest links per paper
  • Up to 3 topic groups
  • Ready in about 10–30 seconds

// three ways in

Start from a topic, a title or a DOI

One search box takes all three. What you type decides how the map is built.

  • A research topic

    Type a question or a few keywords. We run a relevance search on OpenAlex and keep the 40 best matches — papers with an abstract first, because they can be compared most precisely.

  • A paper's title

    Paste the title of a paper you know. The search finds it and the work that sits closest to it, so you see its whole neighbourhood in one picture instead of a long list of results.

  • A DOI

    The map is built around that one paper. It's pinned at the centre and opens first, surrounded by the works OpenAlex lists as related and the papers it cites, ranked by how close they are to it.

// how a map is built

From a search box to a map, in five steps

Nothing is pre-drawn. Every map is built live from real data in the seconds after you press Build map.

  1. 1

    Find the papers

    Your search goes to OpenAlex, the open catalogue of the world's scholarly work. We keep up to 40 papers — each with its title, authors, year, venue, abstract and citation count.

    OpenAlex37 papers found
  2. 2

    Read what each one is about

    SPECTER2, a language model the Allen Institute for AI trained on scientific papers, reads every title and the opening of every abstract and turns it into a fingerprint of its content. Papers about the same idea get close fingerprints, even when they use different words.

    SPECTER2Reading 37 titles and abstracts
  3. 3

    Link the closest matches

    Each paper is joined to the 4 papers whose fingerprints are closest to its own. Every link carries a similarity score — the more alike two papers are, the darker the line between them.

    Links111 links · each paper to its 4 closest
  4. 4

    Let related papers pull together

    The links work like springs: closely related papers pull each other in, unrelated ones push apart, and the map settles into its natural shape. Clusters you can see are clusters in the research.

    LayoutRelated papers pulled together
  5. 5

    Colour the topic groups

    The papers are split into up to three groups by content, each named after the OpenAlex topic most of its papers share. The most-cited papers get the biggest circles.

    Groups3 topic groups · sized by citations
    • Neural Networks and Applications
    • Artificial Intelligence in Healthcare and Education
    • Speech Recognition and Synthesis

// how to read a map

Five things every map tells you at a glance

No chart literacy needed — size, colour, lines and position each carry one meaning.

  • 12 citations1.4K136K

    Circle size

    How often the paper is cited. Size grows on a log scale, so a famous classic doesn't swamp the newer work around it.

  • Colour

    Its topic group. The strip above the map names every group and counts its papers, so you see at once which threads the field splits into.

  • 94% similar38% similar

    Lines

    How similar two papers are. Darker lines join near-twins; faint ones join papers that are related but less alike.

  • bridges both

    Position

    Papers that sit close together are about similar things. A paper between two groups bridges them.

  • The centre of a DOI map

    When you search a DOI, that paper is pinned in the middle and opens first — everything else is arranged around it.

// explore every paper

Click a paper. Get the whole story.

Every circle opens a panel with what you need to decide whether to read it — and where to go next.

Artificial Intelligence in Education

Practical and ethical challenges of large language models in education: A systematic scoping review

Lixiang Yan, Lele Sha, Linxuan Zhao et al. · 2023

British Journal of Educational Technology · 1,453 citations

Open paper(opens in a new tab)

Abstract

Educational technology innovations leveraging large language models (LLMs) have shown the potential to automate the laborious process of generating and analysing textual content. While various innovations have been developed to automate a range of educational tasks (eg, question generation,…

Most similar papers

similarity
  • The essentials

    Title, authors, year, venue, citation count and the abstract, in one panel.

  • Straight to the paper

    One click opens it on its DOI page — or its OpenAlex record when there's no DOI.

  • Its 5 closest papers

    Ranked by similarity with a score on each. Click one and the map jumps to it.

  • Focus on hover

    Point at any circle for a quick card; everything not linked to it fades back.

  • Every paper in a list

    An A–Z list of the whole map, for keyboard and screen-reader users too.

  • Zoom, pan, full screen

    Pinch, scroll or use the buttons — on desktop, tablet and phone.

// who it's for

Made for the first week of any research question

When you don't know the field yet — or need to prove you do.

  • Start a literature review

    See the main threads of a field before you read a single paper, then work through it group by group.

  • Find what you missed

    Map a paper you cite by its DOI and spot close work your reference list doesn't have yet.

  • Spot the classics

    The biggest circles are the most-cited papers — the ones reviewers expect you to know.

  • Write the related-work section

    Map your own topic and see who works closest to you, with each paper's similarity to back your choices.

  • Move into a new field

    Its main topics and landmark papers in one picture, in under a minute.

  • Supervise and teach

    Show a student how a field fits together, not just a reading list of titles.

// plans & limits

Part of the Pro plan

Every plan that includes the Literature Matrix comes with a daily allowance of new maps. You only use it when a map is actually built for you.

New maps today

2 / 20 used

18 of 20 new maps left today · maps someone already built open free

  • A daily allowance of new maps

    The search bar always shows how many you have left. It starts over at midnight UTC.

  • Maps already built open free

    If anyone has already built a map for the same search, it opens instantly and doesn't count toward your day.

  • Nothing found, nothing counted

    A search that finds no papers, fails or is cancelled is handed straight back to your allowance.

// FAQ

Questions, answered

What the Literature Matrix searches, how it compares papers and what it costs.

  • Yes. Every paper comes from OpenAlex, the open catalogue of scholarly work, with its real title, authors, year, venue, abstract and citation count — and a link to its DOI page or OpenAlex record.

// map your field

Your next literature review starts with a map

Search a topic, a title or a DOI and see the research around it in seconds.