Complex Reasoning over Logical Queries on Commonsense Knowledge Graphs

This is the official code and data repository for the ACL 2024 paper: Complex Reasoning over Logical Queries on Commonsense Knowledge Graphs.

1. Dataset

The human-annotated evaluataion set and the distantly-supervised instruction tuning dataset can be found at data.

2. Experiments

• OpenAI LLM prompting experiments

Under the prompting folder.

• Zero-shot QA Model

Under the zero_shot_csqa folder.

Processed evaluated Commonsense QA datasets can be found at here

• Generative commonsense inference

Under the comet folder.

3. Citing this work

@article{fang2024complex,
  title={Complex Reasoning over Logical Queries on Commonsense Knowledge Graphs},
  author={Fang, Tianqing and Chen, Zeming and Song, Yangqiu and Bosselut, Antoine},
  journal={ACL},
  year={2024}
}