BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//programme.europython.eu//europython-2024//speaker//MTEU
 CU
BEGIN:VTIMEZONE
TZID:Europe/Prague
BEGIN:DAYLIGHT
DTSTART:20230712T000000
TZNAME:CEST
TZOFFSETFROM:+0200
TZOFFSETTO:+0200
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20231029T030000
RDATE:20241027T030000
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20240331T030000
RDATE:20250330T030000
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Are LLMs smarter in some languages than others? - Pavel Král
DTSTART;TZID=Europe/Prague:20240711T130000
DTEND;TZID=Europe/Prague:20240711T140000
DTSTAMP:20261014T015406Z
UID:pretalx-europython-2024-KXF8JY@programme.europython.eu
DESCRIPTION:Have you ever asked yourself if Large Language Models (LLMs) p
 erform differently across various languages? I have.\n\nIn this poster ses
 sion\, I will demonstrate how tokens\, embeddings\, and the LLMs themselve
 s perform when utilized in 30 different languages. I will illustrate how l
 anguages influence pricing and various model characteristics.\n\nSpoiler:
 \n- The Greek language is the most expensive to process by most models.\n-
  Processing Asian languages on Gemini is cheaper.\n- You can save up to 15
 % of tokens by removing diacritics.
LOCATION:Main Hall A
URL:https://programme.europython.eu/europython-2024/talk/KXF8JY/
END:VEVENT
END:VCALENDAR