Abstract: A trained machine learning model intercepts a subset of activation values of a generative artificial intelligence inference model executing at a processing system and classifies the subset of activation values as malicious based on the subset of activation values meeting one or more criteria. The trained machine learning model triggers an exception in response to the intercepted subset of activation values being classified as malicious.

Abstract: The invention provides method for producing butanediol and specifically 2,3-butanediol by anaerobic fermentation of a gaseous substrate comprising CO by a culture of at least one micro-organism. In accordance with particular methods of the invention, 2,3-butanediol is produced by anaerobic fermentation of substrates including carbohydrates and carbon monoxide. The invention further provides for the up-regulation of a native 2,3-butanediol dehydrogenase gene in the at least one micro-organism. The 2,3-butanediol can be further converted to compounds such as butene(s), butadiene and methyl ethyl ketone. These compounds can be further converted to chemical products such as 2-butanol.