Code and data for paper Context-faithful Prompting for Large Language Models.
Before you begin, make sure to install the necessary packages: openai, scipy, numpy, tiktoken, tqdm, and scikit-learn. To do so, run the following command:
pip install -r requirements.txt.
Download the NQ and RealtimeQA datasets from Google Drive and extract them to the repository folder. Please note that the TACRED dataset is not included due to its LDC license.
Insert your OpenAI API key to api_secrets.py.
Run experiments on the NQ dataset in the knowledge conflict setting using the following command:
python knowledge_conflict.py --schema ${SCHEMA} --demo_mode ${DEMO_MODE}
To perform experiments on the RealTime QA dataset in the abstention setting, use this command:
python abstention.py --schema ${SCHEMA} --demo_mode ${DEMO_MODE}
The SCHEMA parameter refers to the prompting templates described in the paper and can take the following values: base, attr, instr, opin, or instr+opin. The DEMO_MODE parameter represents the demonstration method, with possible values being none (zero-shot), counter (counterfactual demonstrations, applicable only in the knowledge conflict setting), and original (original demonstrations).
Please be aware that running experiments can be costly. Few-shot evaluation on the full dataset is estimated to cost around $150 for NQ and $30 for RealTime QA when using the text-davinci-003 engine for each prompting templates.
@article{zhou2023context,
title={Context-faithful Prompting for Large Language Models},
author={Zhou, Wenxuan and Zhang, Sheng and Poon, Hoifung and Chen, Muhao},
journal={arXiv preprint arXiv:2303.11315},
year={2023}
}