Research judged on substance, not signal

Mission

PubVerse exists to change how new research is judged. For most of the history of science, a good idea has had to wait for the crowd to catch up before anyone treats it as important. We think the substance of a result can be recognized on the day it appears, and not years later. That belief is the whole point of this work.

PubVerse, alongside fortytwodegrees.org, points the waymark of emerging knowledge and enlightenment. We quantify hidden relationships within knowledge spaces so ideas are evaluated on substance, not signal, and the ones that matter are recognized sooner.

PubVerse is built alongside fortytwodegrees.org and shares its aim of pointing toward emerging knowledge before it becomes obvious. The compass you see turning while PubVerse is at work is the same waymark.

What PubVerse is

PubVerse reads a research abstract and tells you three things about it: how new the idea is, how sound the methods look, and how much the work is likely to matter. It reaches those judgments by weighing the abstract against the published scientific literature in its field, not against a popularity score.

The guiding idea is substance over signal. Signal is everything that surrounds a paper and has nothing to do with whether it is right or important: the name of the journal, the size of the lab, the number of people already crowded around the same fashionable question. Substance is the work itself. PubVerse is built to look past the signal, read the substance, and say what it sees in plain language.

You get numbers, but the numbers are not the product. Each result comes with a short written verdict that explains, in ordinary terms, why the work reads as new, or careful, or likely to matter. The reasoning is meant to be read and argued with, the way you would argue with a thoughtful colleague.

How every judgment is grounded

Every score PubVerse produces is traceable to specific prior work retrieved from the database. When PubVerse says a paper is novel, it cites the nearest prior publications it compared against. When it says the methods are sound, it points to the prior art it used as the benchmark. No judgment comes from a model's internal knowledge alone.

The process works in three steps. First, PubVerse retrieves the most relevant prior work from its corpus of more than 116 million publications using semantic search over the full text of the scientific record. Second, it extracts the substance of each retrieved passage so the comparison is about findings and methods, not keyword overlap. Third, a local language model reads the retrieved evidence alongside the candidate abstract and produces a grounded verdict with citations to the specific prior work behind each claim.

This means every result has receipts. On the Compass page, each needle carries an expandable section showing the prior work it was evaluated against, how close each prior sits in similarity, and a short passage describing what that prior did. An expert can verify or challenge any call in seconds by reading the same evidence the model used.

When PubVerse cannot find relevant prior work in the database, it says so honestly. A paper whose contribution is absent from the retrieved work is not automatically scored as highly novel. PubVerse withholds its judgment rather than inflating a score on an uncontested claim, because the absence of a retrieved pre-emptor does not prove novelty. It may simply mean the relevant prior is not in the corpus or was not retrieved.

Searching for authors, funders and subjects

Search is the first tab, and the box in the top bar reaches it from anywhere (on a phone, it sits behind the magnifier button). Type a name, a funder or a subject. Matches appear as you type, grouped into authors, funders and subjects, and a small slip in the spelling is forgiven. Press Enter for the full list. To find one exact person, open Advanced and paste their ORCID iD, or its orcid.org link: an ORCID iD matches that person and nobody else, however their name is written. Typing an ORCID iD into the main box works the same way.

The publication record behind search is an OpenAlex snapshot with works through March 2026, so a paper or a move after that is not reflected yet, and an author's institution is the one most often listed on their papers of their last three active years. A common name is best searched with an institution or a subject added, such as David Lee Sandia: institution and subject words count toward the match. Subject pages list people who have published since 2022, so the list shows people active now rather than lifetime citation totals alone. OpenAlex sometimes merges two people who share a name into one profile; the profile then shows both people's papers, and its institution can be the other person's.

Searching in one kind

Above the search box, a control reads Authors, Funders, Subjects and All. All, the default, searches everything as before. Choosing Authors, Funders or Subjects looks only for that kind, in the results and in the suggestions under the top-bar box, so a common surname is not buried under funders and topics. The choice is part of the page address, so a link to a results page keeps it.

How authors are ranked

Exact words come first. If you type a name, authors whose name contains every word you typed as a whole word are listed before anyone who matches only by a prefix, a spelling slip or a similar-sounding name. Typos and partial names still work: when no author matches the exact words, the forgiving match fills the list. Among authors that match equally well, the order follows the publication record (citations and h-index), not a score.

Merged profiles

Some records in the bibliographic source merge several people who share a name into one profile, which inflates every count. A profile that looks like that (more than 2,000 works, or more than 150 papers a year) carries a note saying it may merge several people, and it is listed after other matches. A name with a single word, such as a surname alone, is listed after fuller matches unless an ORCID iD is linked. No score is shown for an author, and nothing about an author is part of any novelty, methods or impact score.

Funder counts are a floor

Funding metadata covers about a third of papers, and arXiv preprints carry none. Every funder count in search (results, suggestions, a funder's page and an author's funder chips) is therefore a floor, marked "partial data": the true number of papers a funder supported is higher, and a paper with no funder listed is not an unfunded paper.

Subject pages

Choosing a topic or subfield opens a page for it with the people who publish most in that area, each linking to their own page, and an "Open in Compass" button that carries the subject over to the Compass. Broader levels, domains and fields, open in the Compass directly.

An author's page

The page opens with a short Primary work summary, then the author's map, drawn the same way as the research-space maps: same shapes, colors, legend, panels and search. The author sits at the center. Around them are the themes and subthemes of the papers they lead, and the OpenAlex fields and subfields those papers sit in. Two kinds of people are drawn, and the legend keeps them apart: co-authors have published with the author (the line is thicker the more papers they share), and context authors publish on the same subjects without necessarily having written together; the panel says which subthemes they share. Click the author for the summary and a Highlight menu that dims everything except one field, subfield, theme or subtheme. Click a person for their panel; Expand adds that person's own co-authors and context authors, up to a size cap. Under the map the same information is listed as text: fields and subfields, themes and subthemes, co-authors and context authors. The first time an author is opened it can take up to a minute to build the map.

Exporting an author

Below the map, Export (Excel) downloads a workbook with four sheets: the author, their papers (title, year, DOI, citations, role, corresponding, OpenAlex field, subfield, topic, theme, subtheme), their co-authors (with joint papers and the latest joint date) and their context authors (with the subthemes they share). CSV downloads one of those three lists on its own, and BibTeX downloads their papers as citation entries. Exports need a signed-in account, like the rest of search, and contain no email addresses or social handles.

The page lists all of their papers, most cited first or newest first, with badges for first-author, last-author and corresponding-author papers, the subtheme each paper belongs to, and a link to each DOI. A paper that OpenAlex lists under a different profile is marked "Filed by OpenAlex under another profile". Their most frequent funders link to each funder's page.

Funders

A funder's page shows its other names, how many indexed papers acknowledge it, the subfields it funds most and the authors it funds most. Funding metadata covers about a third of papers, and arXiv preprints carry none, so a paper or a person missing from a funder's page is not evidence that they were unfunded.

Author, funder and citation records come from OpenAlex. The bylines on Compass needles come from the preprint servers, and a name links to an author's page only when the match is unambiguous.

Authors on the research-space maps

A research space is a field PubVerse has mapped: its recent papers are grouped into subthemes, and the subthemes into themes. Each theme and subtheme is linked to the OpenAlex field and subfield its papers sit in. Where a space has a map, its subject page offers Open the map, and an author's page offers Show on the map for each of their subfields that has one.

How author maps are built

Sources. Publication records come from an OpenAlex snapshot (works through March 2026). Nothing is fetched live and no author data is used in any score.

Lead and senior papers. The subjects on an author's map are built from the papers where the author is first author, last author or corresponding author. An author with fewer than three such papers has all of their papers counted. Errata, retractions, paratext and peer-review records are not counted, and a preprint and its published article count once. When an author has more than 150 such papers, the 150 most recent build the subjects and the page says so. The papers list beneath the map shows every paper, including those where the author is a middle author.

Themes and subthemes. Papers are grouped by the similarity of their titles. The small groups are the subthemes and the larger groups they nest in are the themes. Titles and the Primary work summary are written by our local language model from the paper titles and their OpenAlex fields and subfields; the titles name the method or approach and what it is applied to. If the model's answer fails our checks, a plain title taken from the OpenAlex subfield is used instead.

Fixing OpenAlex identities. OpenAlex sometimes splits one person across two profiles or files a paper under a different person with a similar name. Two repairs are applied, only when the evidence is strong, and every repaired record is marked on the page. A co-author listed under two profiles is merged when the names are compatible, the two profiles never appear on the same paper, and they share another collaborator or an institution. A paper of the author's own that OpenAlex filed under a different profile is added when at least two of the author's established collaborators are on it; it is marked "Filed by OpenAlex under another profile".

People filters. A co-author or context author shows only with at least 5 papers in all. A co-author also needs at least 2 joint papers with the author, unless one joint paper is less than 24 months old. Both numbers can be changed under Advanced on the map. The author is never filtered out, and neither is the person you have selected. The text lists beneath the map apply the default numbers, with a Show all button.

Field maps

A field map has two levels: Themes, the big areas, and Subthemes, which opens every theme. OpenAlex fields and subfields appear as their own nodes at both levels, connected to the themes and subthemes whose papers they cover. The search box above the map finds themes, subthemes, OpenAlex fields and subfields, and people. Choosing a person places them on the map with their co-authors and context authors in that field. If the person has no paper in the span the map covers, the box says No recent work in this field and names the span.

Finding people on a map

Click a person and their own co-authors and context authors appear next to them, each with the reason it is shown. Open author page in the person's panel goes to their full page.

Who can see it

Authors on maps are for signed-in accounts, like search. The map pages themselves contain no author data: when you ask for a name, the page fetches it from the PubVerse service after checking you are signed in, and hands it to the map. If you are signed out, or the service is busy, the map says so and shows no authors.

Coverage

Authors are placed from papers that can be matched to the OpenAlex snapshot behind search, which runs through March 2026. Preprints posted after that are on the map as papers but their authors are not placed yet, so a recent author may be missing or look less active than they are. Profiles that look like several people merged under one name are never suggested as neighbors. Authors on maps are shown for orientation only; they are never part of any novelty, methods or impact score.

How to score an abstract

Scoring is the quickest way to see PubVerse at work, and it lives on the landing page, under the tab Course Plot (submit your idea or abstract). You can paste a finished abstract or describe an idea in a few sentences; the box accepts a short idea as well. An abstract that states a hypothesis, the methods and an impact is read most accurately, because each is scored separately.

Paste the abstract into the box. A live counter shows your word count against the limit. If you run over, the counter turns red and the Score button waits until you trim it. Press Score. PubVerse first makes sure the text is a genuine research abstract and quietly turns away posters, fragments, and writing that is not research at all. If the text clears that check, PubVerse evaluates it and returns three measures, each on a scale of 0 to 10: Novelty, Methods, and Impact. Below them sits a short verdict that puts the three numbers into a sentence or two of plain reasoning.

You do not need a title, authors, or a DOI to score an abstract. Those are asked for only if you decide to promote a result. Promoting adds the identifying details, checks that a DOI you supply really belongs to the work, and gives you a link you can share with anyone. Everything you score is kept in your own private history so you can return to it, and you can delete any of it whenever you like.

Reviewing a whole document

Under the paste box on the Course Plot tab you can drop a Word (.docx) or PDF file instead, or press Choose a file. The file can be up to 25 MB, and a PDF up to 60 pages. Older .doc files, password-protected files and files with macros are turned away with a note saying why; save the document as a plain .docx and try again. Dragging works on a computer; on a phone, Choose a file opens your files.

A review takes a few minutes. The page shows each step as it happens (reading the file, checking the submission, reviewing against prior work, placing the comments, writing the document), and says where you are in line when other reviews are ahead of yours. You can leave the page open and come back; a review in progress is picked up again if you reload.

What you get back is a Word document. A .docx comes back as your own document with your text and formatting untouched; a PDF is read first and returned as a Word document built from what was read. Either way the review is added as Word comments in the margin, authored by PubVerse: a summary of the three measures at the top, then comments on the sentences that carry the novelty, the methods and the impact, each with the prior work it was judged against. The same measures, verdict and sources appear on the page, and the file name is kept in your history.

Your uploaded file is deleted as soon as the review ends. The reviewed document stays available to download for 30 days and is then removed; the scores, reasoning and sources stay in your history until you delete them.

How the Compass works for you

The Compass is for watching a whole area of research rather than a single abstract. It takes its name from the 42 compass: a way to keep your bearings while a field moves.

Name a topic you care about. PubVerse confirms that it can cover that area, and tells you plainly if it cannot. Then you choose how far back to look, from the last month out to the last two years. PubVerse sweeps the recent literature in that area and hands back a short list of needles, the results worth your attention, each carried with the same three measures and a plain verdict. It also shows how much material it read to find them, so you can see the size of the haystack behind the short list.

You can save a topic to your profile so it is one click away next time, leave an email address to be told when a sweep has finished, and export a simple list of what the Compass found, either the needles alone or the full scope it searched.

What is a needle

A needle is a result that is both genuinely new and genuinely sound, found while it is still easy to miss.

Most research, by its nature, extends what is already known. That work is valuable, but it is not a needle. A needle breaks new ground, and it holds up when you look at it closely. The hard part is that these two qualities rarely announce themselves early. A truly new result often has few citations, sits in an unfashionable corner, or comes from a group no one is watching yet. By the time the wider field agrees it mattered, the moment to have acted on it early is gone.

The Compass exists to catch needles in that early window, when recognizing them is worth the most. When PubVerse calls something a needle, the verdict says in plain words why: what is new about it, and why the work looks trustworthy rather than merely surprising.

Reading the map

Above the needles is a map of the field itself. It is not a picture of the papers PubVerse judged. It is a picture of what the field has been working on, built from every preprint posted in the window you chose, with your needles placed on it.

The map opens on themes. A theme is a broad area of the field. Each one is a blue circle sized by how many papers it holds, with the number of subthemes inside it written on it, and it carries a short name. Your needles are the red stars, attached to the themes they belong to, so the first thing you see is where in the field your results fall. Gray squares are OpenAlex fields and subfields, connected to the themes and subthemes whose papers they cover.

Opening a theme

Click a theme to open it where it stands. It unfolds into its subthemes, and clicking it again closes it. A subtheme is a group of papers that talk about the same thing. The groups are found by the same code that groups papers inside the scoring pipeline, so a subtheme here means what it means everywhere else in PubVerse, and each one is named by reading its own papers. Clicking a subtheme lists the papers that sit closest to its center, which is the quickest way to check whether a name is fair.

Subthemes come in three kinds, and each has its own shape. Pentagons are problem statements: what the papers say still needs solving. Triangles are outcomes: what the papers report or promise. Hexagons are methods, either computational or experimental. Methods are the busiest layer, so they start hidden. The Methods row in the key at the left switches them on and off.

Themes are a reading aid built on top of the subthemes. PubVerse groups the subthemes that cover the same subject and names each group. A subtheme that is only front matter, or that holds too few papers to place, belongs to no theme and appears only at the Subthemes setting.

Granularity

Granularity is how finely the field is divided, and it has two settings. Themes, the default, shows the big areas and your needles. Subthemes opens every theme at once and adds the subthemes that belong to no theme and the nominated papers (the amber diamonds, papers that scored above the run average). Each step adds more of the field and makes the picture busier without adding information about your own papers.

Drag to move, scroll to zoom while the cursor is over the map, and use Expand or Full screen when you want room to work. Subtheme names appear as you zoom in, and the small (i) beside each needle in the list below opens its theme and flies the map to that paper.

Advanced: the field fit band

Most readers never need the band, so it sits under Advanced at the bottom right of the map, closed. While it is closed, the map applies the default cut it has always used, so nothing changes if you leave it alone. Open it to move the cut yourself. It only affects the nominated papers, which are drawn at the Subthemes setting.

The band asks how far inside the established structure of the field a paper sits, and answers in standard deviations.

Every paper is compared to the cluster it most resembles, which gives a similarity score. That number alone means very little, because clusters differ: some are tight and specific, others broad and loose, so the same score means different things in different parts of the field. So the score is measured against a reference. When the landscape is built, PubVerse records what a random paper from that field scores against each cluster, and how much that varies. A paper's band value is how far above that reference it falls, counted in standard deviations of the reference itself.

A value near zero means a paper looks about as related to its nearest cluster as any paper picked at random would. Two standard deviations means it sits squarely inside recognised work. The slider hides nominated papers below whatever value you choose, so dragging it right leaves the papers most firmly inside the field's existing structure, and dragging it left brings back the ones sitting further out.

Three things the band is not. It is not a model output: no scoring head is involved, and nothing here comes from the novelty model that judges your abstracts. It is arithmetic on where a paper sits relative to its neighbours, and nothing more. It is not a quality score, and it is not novelty. A paper can fit the field closely and be dull, or sit at the edge because it is unusual or because it is simply off topic, and the band cannot tell those apart. We measured the relationship between the band and the judged novelty of a paper and found it weak but real, so it is a soft ordering rather than a test. Treat it as a way to change what the map emphasises, not as a verdict.

Needles are drawn at every setting of the slider. That is a deliberate choice and it is worth understanding, because it means the fact that your needles survive a high cut says nothing at all. They would survive any cut, since the slider never removes them. What the readout tells you instead is how many of the nominated papers are hidden at that setting, which is a real number.

Affiliation