Activation Addition steers language models by adding contrastive activation vectors from prompt pairs to control high-level properties like sentiment and toxicity at inference time without training.
Hopkins, David Bau, Fernanda Viégas, Hanspeter Pfister, and Martin Wattenberg
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
other 1
citation-polarity summary
fields
cs.CL 1years
2023 1verdicts
UNVERDICTED 1roles
other 1polarities
unclear 1representative citing papers
citing papers explorer
-
Steering Language Models With Activation Engineering
Activation Addition steers language models by adding contrastive activation vectors from prompt pairs to control high-level properties like sentiment and toxicity at inference time without training.