© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 127913
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on numeric position metadata—specifically the line‐and‐column numbers or coordinate pairs that annotate code locations.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408

    numbers and code elements

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
     largely
    -0.89
     pérdida
    -0.88
     stemming
    -0.85
     mainly
    -0.85
     not
    -0.85
     darbo
    -0.84
     fuertes
    -0.84
    öszi
    -0.84
    endant
    -0.84
     Dont
    -0.84
    POSITIVE LOGITS
     みた
    1.04
     ещё
    0.89
    كى
    0.89
    ftet
    0.88
    的关系
    0.88
     but
    0.88
    (""));
    0.86
    腩
    0.86
    urte
    0.86
     сне
    0.86
    Activations Density 0.002%

    No Known Activations