© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5194
    Prev
    Next
    INDEX
    Explanations

    The neuron selectively activates on numeric tokens—especially floating‐point or multi‐digit numerical values—flagging numbers and measurements in the text.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     republika
    0.53
    <unused2197>
    0.48
     ampia
    0.47
     galería
    0.45
     republik
    0.45
     rekla
    0.42
     maî
    0.42
     gezegd
    0.42
     accolade
    0.41
    țit
    0.41
    POSITIVE LOGITS
    4
    0.45
    Vol
    0.45
    ©
    0.45
    ${
    0.44
    0
    0.43
    8
    0.42
    Downloaded
    0.42
    6
    0.42
    5
    0.40
    best
    0.40
    Activations Density 0.007%

    No Known Activations