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

    Neuronpedia

    INDEX
    Explanations

    the

    np_max-act-logits · gemini-2.0-flash

    metadata or system-level instructions in text, particularly related to AI language models and configuration details.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @kdearsty

    the

    np_max-act · claude-3-5-haiku-20241022Triggered by @kdearsty

    definite article "the" followed by nouns, especially in question contexts.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @jacktpayne51
    New Auto-Interp
    Negative Logits
    hurt
    -0.09
     nghe
    -0.08
    _it
    -0.08
    fällen
    -0.08
    (delta
    -0.08
    chè
    -0.08
     này
    -0.08
    YO
    -0.07
     την
    -0.07
    την
    -0.07
    POSITIVE LOGITS
     rationale
    0.12
     bedoeling
    0.11
     gist
    0.11
     उद्देश्य
    0.11
     verdict
    0.10
     significance
    0.10
     तरीका
    0.10
     أبرز
    0.10
     takeaway
    0.10
     exactement
    0.10
    Act Density 0.021%

    No Known Activations