Skip to content

Latest commit

 

History

History
4 lines (4 loc) · 460 Bytes

README.md

File metadata and controls

4 lines (4 loc) · 460 Bytes

REASONING LIKE ARISTOTLE: DIAGNOSING AND IMPROVING LANGUAGE MODELS FOR LOGICAL REASONING THROUGH SYLLOGISMS

  • syllo.ipynb contians examples of generating syllogism questions and prompt GPT-3 for answering generated questions.
  • train_comp.py & roberta_comp.py & roberta_n_sample.py contain code for testing RoBERTa with fine-tuning.
  • train_folio.py & train_folio_base.py & zero_folio.py contain code for testing RoBERTa on FOLIO with/without pre-training.