© 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. 5969
    Prev
    Next
    INDEX
    Explanations

    This neuron specializes in spotting numeric dimension or resolution specifications (e.g. “WxH,” “3×3,” “pixels,” “array”) in the text.

    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
    馏
    0.63
     NumericVector
    0.60
     Athletes
    0.54
     nightclub
    0.53
    摄
    0.53
    美元
    0.53
     нарко
    0.52
    怂
    0.52
     Athlete
    0.51
    忽
    0.51
    POSITIVE LOGITS
     grid
    1.70
     grids
    1.66
    grid
    1.58
    GridLayout
    1.48
     Grid
    1.42
     matrix
    1.41
    Grid
    1.38
     GridLayout
    1.38
    GRID
    1.36
     gridSize
    1.35
    Activations Density 0.228%

    No Known Activations