Enhancing Large Language Models Against Inductive Instructions with Dual-critique Prompting

Opensource

We have reorganized INDust, aligning evidence with three types of inductive instructions and implementing stricter quality control measures.

• INDust & LINDust.
• Evaluation prompts and parameter settings are already presented in the paper.
• Code of Finetuning LLMs on LINDust.

Citations

@online{Wang2023b,
  title = {Self-Critique Prompting with Large Language Models for Inductive Instructions},
  author = {Wang, Rui and Wang, Hongru and Mi, Fei and Chen, Yi and Xu, Ruifeng and Wong, Kam-Fai},
  doi = {10.48550/arXiv.2305.13733},
  url = {http://arxiv.org/abs/2305.13733},
  urldate = {2023-05-26},
  pubstate = {preprint}
}