Prompt Compression and Contrastive Conditioning for Controllability and Toxicity Reduction in Language Models. (arXiv:2210.03162v1 [cs.CL])

We explore the idea of compressing the prompts used to condition language
models, and show that compressed prompts can retain a substantive amount of
information about the original prompt. For severely compressed prompts, while
fine-grained information is lost, abstract information and general sentiments
can be retained with surprisingly few parameters, which can be useful in the
context of decode-time algorithms for controllability and toxicity reduction.
We explore contrastive conditioning to steer language model generation towards
desirable text and away from …

arxiv compression language language models toxicity