« COPEN » : différence entre les versions
(Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' COPEN''' == Anglais == '''COPEN''' A new benchmark evaluates PLMs' conceptual knowledge through tasks assessing conceptual similarities, properties, and context comprehension, revealing significant gaps compared to human-like cognition. == Source == [https://huggingface.co/papers/2211.04079 Source : huggingface] Catégorie:vocabulary ») |
Aucun résumé des modifications |
||
| Ligne 10 : | Ligne 10 : | ||
'''COPEN''' | '''COPEN''' | ||
A new benchmark evaluates PLMs' conceptual knowledge through tasks assessing conceptual similarities, properties, and context comprehension, revealing significant gaps compared to human-like cognition. | A new benchmark evaluates PLMs' conceptual knowledge through tasks assessing conceptual similarities, properties, and context comprehension, revealing significant gaps compared to human-like cognition. | ||
Probing Conceptual Knowledge in Pre-trained Language Models | |||
Version du 27 mai 2025 à 10:30
en construction
Définition
XXXXXXXXX
Français
COPEN
Anglais
COPEN
A new benchmark evaluates PLMs' conceptual knowledge through tasks assessing conceptual similarities, properties, and context comprehension, revealing significant gaps compared to human-like cognition.
Probing Conceptual Knowledge in Pre-trained Language Models
Source
