ELOQUENT

The ELOQUENT lab for evaluation of generative language model quality and usefulness addresses high-level quality criteria for generative language models through a set of open-ended shared tasks.

Tasks

The ELOQUENT Lab features three different tasks:

Voight-Kampff
Can machine-generated text be distinguished from human-authored text?
Robustness
Will a generative language model’s output reflect cultural variety? Will it be able to provide robust responses irrespective of interaction language?
Topical PISA Quiz
Can a generative language model create a useful topical quiz from given text? Can it score responses to such quiz questions?

Organizers

  • Jussi Karlgren (AMD Silo AI)
  • Marie Isabel Engels (Fraunhofer IAIS)
  • Maria Barrett
  • Diandra Fabre
  • Pavel Šindelář (Charles University)
  • Ondřej Bojar (Charles University in Prague, ÚFAL)
  • Lorraine Goeuriot
  • Josiane Mothe
  • Philippe Mulhem
  • Mario Piacentini (OECD)
  • Luis Francisco Vargas Madriz (OECD)
  • Didier Schwab
  • Georgios Stampoulidis
  • Katherina Thomas (OECD)
  • Markarit Vartampetian