DIRECTOR: Generator-Classifiers For Supervised Language Modeling. (arXiv:2206.07694v1 [cs.CL])

Web: http://arxiv.org/abs/2206.07694

Current language models achieve low perplexity but their resulting
generations still suffer from toxic responses, repetitiveness and
contradictions. The standard language modeling setup fails to address these
issues. In this paper, we introduce a new architecture, {\sc Director}, that
consists of a unified generator-classifier with both a language modeling and a
classification head for each output token. Training is conducted jointly using
both standard language modeling data, and data labeled with desirable and
undesirable sequences. Experiments in several settings show …

arxiv generator language modeling