Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    pronouns and sentence beginnings

    np_acts-logits-general · gemini-2.5-flash-lite

    scripted multi-speaker dialogue structure, especially speaker turn labels and direct-address conversational turns typical of roleplay or staged conversations.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Negative Logits
     automaton
    0.29
     appendages
    0.29
     quarrels
    0.29
     stylesheets
    0.29
     palettes
    0.28
     paradigms
    0.28
     vases
    0.28
     strands
    0.27
     qubit
    0.27
     rune
    0.27
    POSITIVE LOGITS
    you
    0.43
    we
    0.39
     we
    0.37
     আমি
    0.37
     I
    0.36
    আমি
    0.36
    我們
    0.36
     मैं
    0.35
    আমরা
    0.34
    मैं
    0.34
    Act Density 0.562%

    No Known Activations