Bo
Boolean Neurons
Neurons encoding truth values, polarity, and binary distinctions.
About Boolean Neurons
Boolean neurons are the gatekeepers of AI reasoning. They encode the most fundamental binary operations a language model must perform: is something true or false, present or absent, affirmed or denied? Boolean neurons appear across all model architectures we tested, though they represent a small fraction of the total — roughly 1–2 percent of classified features. Their scarcity is deceptive: Boolean signals propagate throughout the network, modulating how other neuron types behave. A suppressed Boolean.Affirmation neuron in an early layer can cascade into incorrect factual outputs ten layers later.
The 7 Boolean Elements
1 Bo
Affirmation
Encodes positive confirmation and agreement signals.
Explore →
2 Bo
Negation
Encodes denial, refusal, and negative polarity.
Explore →
3 Bo
Polarity
Discriminates positive vs. negative sentiment and evaluative direction.
Explore →
4 Bo
Presence
Detects existence and presence of entities and states.
Explore →
5 Bo
Binary
Encodes yes/no, on/off, and mutually exclusive binary choices.
Explore →
6 Bo
Comparison
Compares two states and selects a winner or identifies their relationship.
Explore →
7 Bo
General
Boolean neurons not fitting a specific sub-type.
Explore →