© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 603
    Prev
    Next
    INDEX
    Explanations

    followed by determiners

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

    The neuron fires on short function words that introduce spatial or temporal phrases—especially common prepositions (“in,” “on,” “from,” “into,” etc.) and their adjoining articles.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    prepositions and articles in descriptive or narrative contexts, particularly in phrases indicating location, time, or relationships between entities.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    function words—especially prepositions and articles—that link clauses and indicate relations such as location, direction, or association.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    +
    0.50
    /
    0.43
    '
    0.42
    ist
    0.42
    s
    0.40
    '(
    0.39
    _
    0.39
    ies
    0.39
    =
    0.39
    ',
    0.39
    POSITIVE LOGITS
     früheren
    0.42
     său
    0.40
     această
    0.39
     منطقة
    0.39
     publicó
    0.38
     lhe
    0.37
    この
    0.37
     kawasan
    0.37
     około
    0.36
     sogenannten
    0.36
    Activations Density 4.668%

    No Known Activations