How much does an LM depend on information provided in-context vs its prior knowledge?

Check out how Vésteinn Snæbjarnarson, Niklas Stoehr, Jennifer White, Aaron Schein, Ryan David Cotterell + I answer this by measuring a *context's persuasiveness* and an *entity's susceptibility*🧵

Excited to finally have Alexander Hoyle visiting us at ETH AI Center and ETH D-GESS sharing his cool work on „Social Science as a Problem Space for NLP“, co-hosted by Ryan David Cotterell and Elliott Ash.

If you are at #EACL2024 , on Friday 4pm Malta time, I will give a (virtual) talk on 𝗖𝗼𝗻𝘁𝗿𝗼𝗹𝗹𝗲𝗱 𝗗𝗲𝗰𝗼𝗱𝗶𝗻𝗴 𝗳𝗿𝗼𝗺 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 at the 1𝘴𝘵 𝘗𝘦𝘳𝘴𝘰𝘯𝘢𝘭𝘪𝘻𝘢𝘵𝘪𝘰𝘯 𝘰𝘧 𝘎𝘦𝘯𝘦𝘳𝘢𝘵𝘪𝘷𝘦 𝘈𝘐 (𝘗𝘌𝘙𝘚𝘖𝘕𝘈𝘓𝘐𝘡𝘌) 𝘸𝘰𝘳𝘬𝘴𝘩𝘰𝘱.

In today's Oral Presentation session at #EACL2024 (11:30 CET), ETH Zurich's Niklas Stoehr will present 'Unsupervised Contrast-Consistent Ranking with Language Models,' a paper co-authored w/ our #AI researchers Pengxiang Cheng, Jing Wang, Daniel Preotiuc & Rajarshi Bhowmik
#NLProc

Congratulations to Niklas Stoehr of ETH Zurich who co-authored #EACL2024 paper 'Unsupervised Contrast-Consistent Ranking with Language Models' together with our #AI researchers Pengxiang Cheng, Jing Wang, Daniel Preotiuc & Rajarshi Bhowmik
bloom.bg/3VnM0DA
#NLProc
(1/2)

🏛📣 Life Update 📣🏛 This fall, I’ll be starting a position as an Assistant Professor of Computational Linguistics at Georgetown University University! I’m excited to be moving to DC🇺🇸 and to be joining many wonderful colleagues at Georgetown Linguistics

Happy to share the good news that as of today I’m officially tenured — aka promoted from ass. prof. to ass. prof. 🤔
ethrat.ch/en/appointment…

Many thanks to all those who have been supporting me along this exciting path: students, colleagues, and most of all my lovely family. 💋

My ex EPFL lab mates Kristina Gligorić and Tiziano Piccardi invited me to lead a session on “Making Measurements with LLMs” in their Social NLP reading group tomorrow: sites.google.com/view/social-nlp 😍 I’ll present arxiv.org/pdf/2309.06991…

Let me know if you are Stanford University and free to chat.

Preprint for this work (led by sean) is now online: arxiv.org/pdf/2312.09203…

Feedback very welcome!

🚨New round of recruiting! Join our growing team to bridge ML models with interpretable domain expertise.
Learn more: neuroexplicit.org
Apply by Jan 7

Thank you to #EMNLP2023 chairs for the 😱 two 😱 outstanding paper awards! I am so grateful to have worked on these projects with wonderful colleagues — Tiago Pimentel (who is the first author on one of the papers!), Clara Isabel Meister, Kyle Mahowald and Ryan David Cotterell

… just so sad to miss this amazing program at the BlackboxNLP workshop at #EMNLP2023 ! Invited talks by Zhijing Jin and Antoine Bosselut (hop 🇨🇭!) and a panel discussion on Mechanistic Interpretability involving Neel Nanda, John Hewitt and Yonatan Belinkov. ◼️🔍

Our #AI Engineering Group' s Sr. Research Scientist Daniel Preotiuc will give an Industry Keynote on 'Modular Language Modeling through Model Merging' at today's 2023 Singapore Symposium on Natural Language Processing (16:30 SGT)
bloom.bg/47INeN1
#nlproc #EMNLP2023 #SSNLP

My US stint continues: I am super excited to spend the next three months at Google in the Bay Area working on Language Model Interpretability GoogleStudents

🔵🔴🟡🔵🟢🔴