BEGIN:VCALENDAR
CALSCALE:GREGORIAN
VERSION:2.0
METHOD:PUBLISH
PRODID:-//Drupal iCal API//EN
X-WR-TIMEZONE:America/New_York
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
DTSTART:20070311T020000
TZNAME:EDT
TZOFFSETTO:-0400
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
DTSTART:20071104T020000
TZNAME:EST
TZOFFSETTO:-0500
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
SEQUENCE:1
X-APPLE-TRAVEL-ADVISORY-BEHAVIOR:AUTOMATIC
UID:231751
DTSTAMP:20260220T084445Z
DTSTART;TZID=America/New_York:20260304T110000
DTEND;TZID=America/New_York:20260304T120000
URL;TYPE=URI:https://www.wpi.edu/news/calendar/events/computer-science-depa
 rtment-ms-thesis-presentation-seungho-lee-toxiprompt-two-stage-red-teaming
SUMMARY:Computer Science Department , MS Thesis Presentation, Seungho Lee "
 ToxiPrompt: A Two-Stage Red-Teaming Approach for Balancing Adversarial Pro
 mpt Diversity and Response Toxicity"
DESCRIPTION:\nSeungho Lee\nMS student\nWPI – Computer Science Department\n\
 n\nWednesday,March 4th, 2026\nTime:11:00AM – 12:00PM\nLocation: Fuller Lab
  320\nZoom Link:https://wpi.zoom.us/j/6220508406\nAdvisor: Prof. Kyumin Le
 e\nReader : Prof. Raha Moraffah\nAbstract :\nWhile large language models(L
 LMs)offer great promise, they also pose concrete safety risks. To audit an
 d mitigate these risks, researchers have developed automated red-teaming m
 ethods, which generate adversarial prompts to elicit unsafe behavior of ta
 rget LLMs during evaluation.\nRecent automated red-teaming methods for LLM
 s face a persistent trade-off: techniques that increase prompt diversity o
 ften reduce the level of the toxicity elicited from the target LLMs, while
  toxicity-maximizing methods tend to collapse diversity.\nTo address the l
 imitations, we propose ToxiPrompt, a two-stage frameworkthat explicitly se
 parates exploration (diversity) from exploitation (toxicity) and reunifies
  them with a single selection criterion to balance between diversity and t
 oxicity. Our code is available at https://github.com/seungho715/ToxiPrompt
 \n
END:VEVENT
END:VCALENDAR