BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Memorization as a Feature\, Not a Bug - Jing Huang (Stanford Unive
 rsity)
DTSTART:20260206T160000Z
DTEND:20260206T170000Z
UID:TALK242716@talks.cam.ac.uk
CONTACT:Suchir Salhan
DESCRIPTION:Memorization in LLMs has long been perceived as undesirable\, 
 associated with privacy risks\, copyright concerns\, and wasted capacity. 
 In this talk\, I argue for a complementary perspective: memorization is an
  intrinsic property of LLMs that can be leveraged to build a better LLM ec
 osystem. I first present two frameworks to rigorously study counterfactual
  memorization of a training run. I then demonstrate how memorization dynam
 ics can be exploited to establish model and text provenance. Together\, th
 ese results suggest a new perspective: rather than focusing on suppressing
  memorization\, we should aim to understand and harness it. Doing so opens
  new avenues for provenance\, tracing downstream impacts\, and policies ar
 ound intellectual property and integrity in the LLM ecosystem.\n\n**Speake
 r Bio:** Jing Huang is a PhD candidate in the StanfordNLP Group\, advised 
 by Prof. Christopher Potts and Dr Diyi Yang. Jing's research interests foc
 us on understanding what makes neural network models generalize well by st
 udying the causal mechanisms that connect model behaviors\, internal repre
 sentations\, and training data.
LOCATION:ONLY ONLY. Here is the Google Meet Link: https://meet.google.com/
 cru-hcuo-rhu
END:VEVENT
END:VCALENDAR