© 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
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 2947
    Prev
    Next
    INDEX
    Explanations

    start of turn: model

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

    This neuron detects formal causal or directional prepositional phrases (e.g. “for the reason that of,” “in direction of”).

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

    low-quality, spammy, or machine-generated text with awkward phrasing and formatting.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jrosseruk
    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
     וי
    0.73
    יי
    0.72
    นี้
    0.70
    ~
    0.70
    0.69
     of
    0.68
     столь
    0.66
     have
    0.66
     had
    0.66
     जानेवारी
    0.65
    POSITIVE LOGITS
    attamente
    0.86
     Metadata
    0.85
    itabbo
    0.82
    <unused99>
    0.82
    atthena
    0.81
    <unused481>
    0.81
    िकारक
    0.81
    گنڈ
    0.80
     ArrayRef
    0.80
     Locale
    0.79
    Activations Density 0.056%

    No Known Activations