<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div class=""><span class=""><i class="">Department of Computer Science Seminar</i></span></div>
<div class=""><span class=""><i class=""><br class="">
</i></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">Ari Holtzman</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">PhD Candidate, Computer Science & Engineering </b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">University of Washington</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class=""><br class="">
</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">Tuesday, February 14th</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">2:00pm - 3:00pm</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">In Person: John Crerar Library 390</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class=""><br class="">
</b></span></div>
<div class=""><span style="font-style: normal;" class=""><b class="">Zoom: </b></span></div>
<div class=""><b class="">
<div class=""><a href="https://uchicagogroup.zoom.us/j/91667606873?pwd=UmYvSWNGdm1ES3UwZExicG04Ri9mZz09" class="">https://uchicagogroup.zoom.us/j/91667606873?pwd=UmYvSWNGdm1ES3UwZExicG04Ri9mZz09</a></div>
<div class=""><br class="">
</div>
<div class="">Meeting ID: 916 6760 6873</div>
<div class="">Passcode: 323372</div>
</b></div>
<span class=""><b class="">
<div class=""><span class=""><b class=""><br class="">
</b></span></div>
Title: Controlling Large Language Models: Generating (Useful) Text from Models We Don’t Fully Understand</b></span><span class=""><br class="">
<b class="">Abstract: </b><br class="">
</span><span class="">Generative language models have recently exploded in popularity, with services such as ChatGPT deployed to millions of users. These neural models are fascinating, useful, and incredibly mysterious: rather than designing what we want them
to do, we nudge them in the right direction and must discover what they are capable of. But how can we rely on such inscrutable systems?<br class="">
<br class="">
</span><span class="">This talk will describe a number of key characteristics we want from generative models of text, such as coherence and correctness, and show how we can design algorithms to more reliably generate text with these properties. We will also
highlight some of the challenges of using such models, including the need to discover and name new and often unexpected emergent behavior. Finally, we will discuss the implications this has for the grand challenge of understanding models at a level where we
can safely control their behavior.<br class="">
<br class="">
</span><span class=""><b class="">Bio: </b><br class="">
</span><span class="">Ari Holtzman is a PhD student at the University of Washington. His research has focused broadly on generative models of text: how we can use them and how can we understand them better. His research interests have spanned everything from
dialogue, including winning the first Amazon Alexa Prize in 2017, to fundamental research on text generation, such as proposing Nucleus Sampling, a decoding algorithm used broadly in deployed systems such as the GPT-3 API and academic research. Ari completed
an interdisciplinary degree at NYU combining Computer Science and the Philosophy of Language.<br class="">
<br class="">
</span><img apple-inline="yes" id="655D86DD-F49C-475E-A870-4CA625FF2FA4" src="cid:65F15779-5CCC-44FA-9B07-E794F520FDF0" class=""><span class=""><br class="">
</span><span class="">---<br class="">
</span>
<div class="">
<div>Holly Santos<br class="">
Executive Assistant to Michael J. Franklin, Chairman<br class="">
Department of Computer Science<br class="">
The University of Chicago<br class="">
5730 S Ellis Ave-217 Chicago, IL 60637<br class="">
P: 773-834-8977<br class="">
<a href="mailto:hsantos@uchicago.edu" class="">hsantos@uchicago.edu</a></div>
<div class=""><br class="">
</div>
<br class="Apple-interchange-newline">
</div>
<br class="">
</body>
</html>