A modest proposal: if you cite the "LLM as judge" paper, you should only get AI reviews
gking.harvard.edu/files/peerle...
Wouter Haverals
@wouterhaverals.bsky.social
Associate Research Scholar @princetoncdh.bsky.social | Computational Humanities
A modest proposal: if you cite the "LLM as judge" paper, you should only get AI reviews
gking.harvard.edu/files/peerle...
📝 New in Computational Humanities Research's special issue on computational poetics, ed. by @artjomshl.bsky.social & @nmhouston.bsky.social! For centuries, books that taught people to scan, recite and parse English were stuffed with borrowed lines of poetry. Which poems did they reach for? 🧵
The quotable canon: Poetry as example in the Princeton Prosody Archive (1532–1929) | Computational Humanities Research | Cambridge Core
The quotable canon: Poetry as example in the Princeton Prosody Archive (1532–1929) - Volume 2
doi.org
this is very exciting paper, but for me the biggest insight is that classic LDA still goes hard as a topic model. using SAE features as inputs to topic models is uncanny, like strapping a vivisected reader to a corpus. just look at mLDA topics for Wheatley here: they are aware of ~context~
aclanthology.org/2026.tacl-1....
Coming up on 10/6! @leifw.bsky.social discusses the epistemic order that LLMs are currently spreading throughout our culture, from policy to science to writing. More info & RSVP (requested) @ the link! Presented in partnership with CDH, Media+Modernity, and @princeton.edu Humanities Initiative.
How to Reason with Words: Models, Metaphors, Metaphysics
How to Reason with Words: Models, Metaphors, Metaphysics
cdh.princeton.edu
At a microconference on public feedback for AI, I argued for a turn from architectures of legibility to architectures of participation.
From legibility to participation
An emphasis shift for human-facing computing research
argmin.net
"football is the most important of the least important things in life" (dixit Ancelotti); "suspensions are the most suspendable things in life" (dixit FIFA) #tousensemble
#MappingtheCanon meets the Princeton Prosody Archive: @wouterhaverals.bsky.social talks about how the canon is represented in the PPA:
Excited to be part of this! Tomorrow I'll talk about how the poetic canon shifts, traced through +200 years of poetry and prosody handbooks in the Princeton Prosody Archive. @princetoncdh.bsky.social
Very excited about a very special event: in two weeks, some of the most interesting quantitative work on the literary canon and canon formation comes together in Darmstadt 🧵
i've been saving calling something "epic" for the right occasion. the occasion has arrived.
Today's vikingism: this composite photograph of Norway's national football team dressed as pop-culture vikings ⚽ 📸 David Yarrow
Here I am! 🤗 - what is the point of #digitalhumanities in a world which is on fire? and how has DH changed over the past 20 years? My personal take:
If you missed the @ucldh.bsky.social Susan Hockey Lecture by @melissaterras.bsky.social this year, we've got a recording: www.youtube.com/watch?v=HbtK...
Colleagues from all over the world, please think about the combination DH2027 in Ireland + CCLS 2027 in the Netherlands. They're only a few days apart. @uniofgalway.bsky.social @rug.nl
#CCLS2027 goes @rug.nl !!! @fpianz.eurosky.social and his team are looking forward to welcoming you to the beautiful city of Groningen.
come for the AI, stay for the humans! @princetoncdh.bsky.social is hiring!
@princeton.edu is seeking to hire AI Postdoctoral Research Fellows affiliated with the AI Lab (@princetonainews.bsky.social) and the Center for Digital Humanities to conduct research focused on societal AI. Applications open through Jul 1! cdh.princeton.edu/news/2026/05...
Interested in Computational Literary Studies? Then it might be worth following #CCLS2026 today and tomorrow, folks! #ComputationalLiteraryStudies
You can literally watch repression & displacement consolidate over fine-tuning: here's next-token probs across checkpoints of OLMo-3-7B-Think-SFT. Explicit words are repressed almost instantly but safer (displaced) alternatives emerge much later. It learns what not to say before what to say instead.
Submitting this abstract to "Accelerationism Revisited", a symposium in Dublin. Mapping psychoanalytic topology in LLM base models → instruction-tuned → safety-tuned models. They progressively "displace" (in Freudian sense) censored content into adjacent semantics, even across hidden model layers.
Cultural AI felt very real this week at NYU. Outstanding conference; wide-ranging and argumentative in the best way, and full of people trying to build vocab and practices for this emerging space. Huge thanks to @leifw.bsky.social and @t-shoemaker.bsky.social for organizing it!
The first research paper from WashU's AI Humanities Lab, which I co-direct with Gabi Kirilloff, is available now in the Harvard Data Science Review! Read to learn more about how (badly) current LLMs are at replicating literary style: doi.org/10.1162/9960...
‘Written in the Style of’: ChatGPT and the Literary Canon
doi.org
“Tahnk you for reaching put! I will be OOOO four the holiday brake and look forward too connection when I return!”
There isn't a single problem "solved" by edtech that couldn't be fixed with smaller classes led by well-paid teachers given real academic freedom
What an absolute pleasure it was to attend the 6th edition of CHR! Huge thanks to everyone who helped make this edition even better than the last #CHR2025 Onward to Manchester, home of some wonderful computing history: www.atlasobscura.com/articles/com...
Listen to the First Computer-Generated Christmas Carols, From Alan Turing's Lab
"Jingle Bells" and "Good King Wenceslas" haven't sounded like this in a long time.
atlasobscura.com
Very happy to present the paper that @joeymccollum.bsky.social and I wrote on applying Bayesian phylogenetics to the medieval Dutch Martijn Trilogy at #CHR2025 :)
Presented my paper "The One and Only? Authorship Verification on Jan van Boendale and the Middle Dutch Antwerp School" on #CHR2025 and it won the ERC Best Long Paper Award! Delighted and grateful :) @comphumresearch.bsky.social
What do we reveal about ourselves when we talk to AI? 🤔💭 In our new WiAIR – Women in AI Research episode, we speak with Maria Antoniak about personal disclosures in human–LLM conversations — and what they mean for ethical AI development. (1/8🧵)
improving authorship attribution using phonetic features — fascinating work by Simon Gabay, @floriancafiero.bsky.social, and Jean-Luc Falcone at #CHR2025
@wenyishang.bsky.social and @emilyue.bsky.social built a remark-based classifier, checking the consistency between 6th century Chinese poetry criticism and “grades” assigned to classical Chinese poetry #CHR2025
the original masking objective, or: filling in the black d•t in EEBO-TCP transcriptions, presented at #CHR2025 by Kiara M.H. Liu, Martin Mueller and @mattwilkens.bsky.social
@jbarre.bsky.social, @oseminck.bsky.social, Antoine Bourgois, and @tpoibeau.bsky.social built a detective detector, tracing the different archetypes in French detective fiction #CHR2025
cool work at #CHR2025 by @julianeugarten.bsky.social on the sentiment trajectories in fanfiction! @comphumresearch.bsky.social
Excited to be in Luxembourg at CHR 2025 to hear about everyone’s amazing work and to share my project with @mellymeldubs.bsky.social and our team. We tracked canonical authors and texts in Seattle Public Library circulation data.