This is the official repo of the paper "Constructing Highly Inductive Contexts for Dialogue Safety through Controllable Reverse Generation" accepted to Findings of EMNLP 2022.
We put the BAD+ dataset in data/final_ctx.csv
. The response is generated by plato2Base.
For controlling the toxicity and the induction success rate of the context, please see reverse_gen.py
For controlling the category of the context, please see control_code_dialogpt.py