© 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. 31-GEMMASCOPE-2-RES-16K
    7. 5651
    Prev
    Next
    INDEX
    Explanations

    open/close state

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

    The neuron fires on words that describe opening or closing actions (e.g. “open,” “opened,” “shut,” “closed,” “down”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_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
    דר
    0.54
    屑
    0.50
     jost
    0.49
     jut
    0.48
    MaxLength
    0.48
    ructuring
    0.48
    क्यात
    0.48
    첸
    0.48
     bhith
    0.48
     நிர்
    0.47
    POSITIVE LOGITS
     closed
    2.42
     закры
    2.39
     opened
    2.36
    closed
    2.31
     Closed
    2.31
    閉
    2.31
     CLOSED
    2.28
    opened
    2.25
    Closed
    2.25
     closing
    2.22
    Activations Density 0.449%

    No Known Activations