This neuron detects sentence beginnings, particularly detecting "It is" or "This is" at the start of sentences and paragraphs.

first-person pronouns and helping verbs (particularly "I" and auxiliary verbs like "would", "was", "need", "can").

references to giving preference, priority, or favor to something or someone.

specific abbreviations or acronyms formatted with uppercase letters and numbers, often appearing in technical or legal contexts.

technical acronyms and short codes, particularly those related to scientific and computing domains (like LC, LF, etc.).

text related to theater, opera, drama, and performing arts, especially in non-English languages.

descriptions of furniture and home goods with their properties, materials, and quality features.

The neuron detects the word "gate" in electronic and computing contexts.

phrases related to political history, citizenship rights, and systemic injustices, particularly focusing on complex historical narratives about oppression, rights, and national struggles.

strings and other programming language related terms.

technical keywords, variables, codes, and structured identifiers, especially those related to programming, technical specifications, or data labels.

words relating to Christianity and spiritual practice.

references to Christian faith, church activities, and religious community.

a mix of dates, times, years and code related terms.

technical jargon, codes, or identifiers frequently used in programming, system logs, error messages, and system or network operations.

capitalized text, particularly in titles, headings, and code/programming identifiers.

instances where information or content is made publicly available.

This neuron detects the presence of a question—marking interrogative constructions (e.g. the “Q:” prompt, question‐word phrases, question marks, etc.).

scientific publications, with a particular focus on research methodology, results, and experimental design within those publications.

terms and phrases referring to categories, groups, types, classes, or roles that organize or distinguish entities, especially within technical or scientific contexts.