Aaron Schein
@AaronSchein
Asst Prof @DSI_UChicago & Stats l Previously: postdoc @DataSciColumbia, PhD @manningcics | research on ML, applied stats, social science
ID:246427772
http://www.aaronschein.com 02-02-2011 19:32:28
2,3K Tweets
1,6K Followers
1,2K Following
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*🧵
Does this make LLMs biased? Idk maybe. But here's our new paper on when PageRank (dev by Larry Page and Sergey Brin) and other ranking algs are indeed biased! ➡️arxiv.org/abs/2402.13787
Accepted at The Web Conference'24, who else is joining in Singapore this May?
📢 Excited to deliver an oral presentation at #NeurIPS2023 Workshop on CRL tomorrow!
📄 The Linear Representation Hypothesis and the Geometry of Large Language Models (arxiv.org/abs/2311.03658)
🕒 09:50 am - 10:05 am (Poster: 10:30 am - 12:00 pm)
📍 Room 243 - 245
Excited to reunite with everyone at
#NeurIPS2023 next month! Don't forget to check out our workshop I Can't Believe It's Not Better! on the failure modes of foundation models: shorturl.at/qrtJM (Looks like we added one more foundation model failure mode after last weekend!) 🤓
Did you know that The University of Chicago just launched a new PhD in #DataScience and #AI ? Join the info session on 11/16 to learn more about this new program!
The University of ChicagoPSD datascience.uchicago.edu/events/phd-in-…