COPEN: Probing Conceptual Knowledge in Pre-trained Language Models.
Hao PengXiaozhi WangShengding HuHailong JinLei HouJuanzi LiZhiyuan LiuQun LiuPublished in: CoRR (2022)
Keyphrases
- language model
- conceptual knowledge
- pre trained
- qualitative reasoning
- knowledge structures
- training data
- procedural knowledge
- speech recognition
- information retrieval
- training examples
- conceptual structures
- probabilistic model
- test collection
- control signals
- knowledge construction
- conceptual understanding
- data sets
- dynamic systems
- information extraction
- active learning
- learning algorithm