Extending Activation Steering to Broad Skills and Multiple Behaviours

Autor:	van der Weij, Teun, Poesio, Massimo, Schoots, Nandi
Rok vydání:	2024
Předmět:	Computer Science - Machine Learning Computer Science - Artificial Intelligence Computer Science - Computation and Language Computer Science - Computers and Society
Druh dokumentu:	Working Paper
Popis:	Current large language models have dangerous capabilities, which are likely to become more problematic in the future. Activation steering techniques can be used to reduce risks from these capabilities. In this paper, we investigate the efficacy of activation steering for broad skills and multiple behaviours. First, by comparing the effects of reducing performance on general coding ability and Python-specific ability, we find that steering broader skills is competitive to steering narrower skills. Second, we steer models to become more or less myopic and wealth-seeking, among other behaviours. In our experiments, combining steering vectors for multiple different behaviours into one steering vector is largely unsuccessful. On the other hand, injecting individual steering vectors at different places in a model simultaneously is promising. Comment: Code is available at: https://github.com/TeunvdWeij/extending-activation-addition
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2403.05767 Zobrazit plný text záznamu View this record from Arxiv