© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 45184
    Prev
    Next
    INDEX
    Explanations

    code after strings

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

    The neuron is detecting single-quote characters (the apostrophe tokens) that mark string delimiters in code.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     "
    -1.80
     $
    -1.47
    THEY
    -1.46
    titial
    -1.45
    所谓
    -1.41
    geries
    -1.38
     didn
    -1.33
    цкий
    -1.31
    nu
    -1.31
    self
    -1.30
    POSITIVE LOGITS
     this
    1.91
     selben
    1.84
    ثم
    1.81
     Autó
    1.72
     fuertemente
    1.72
     carico
    1.72
     then
    1.72
     acompañada
    1.69
     тобой
    1.69
     extremadamente
    1.68
    Activations Density 0.007%

    No Known Activations