© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 130629
    Prev
    Next
    INDEX
    Explanations

    The neuron spikes on code‐related tokens that carry numeric indices or chunk labels (e.g. variable names or parts like “part0,” “x,” “_,” or literal digits).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    code examples

    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
    Not in Any Lists

    No Comments

    Negative Logits
    کا
    -1.01
     provar
    -0.93
    ולות
    -0.91
     tiens
    -0.91
     RACING
    -0.90
     pyta
    -0.88
     plafon
    -0.87
     sportif
    -0.86
    ajaja
    -0.85
     Racism
    -0.85
    POSITIVE LOGITS
     from
    0.88
     like
    0.77
     !!
    0.75
     Depending
    0.75
     depending
    0.75
     always
    0.74
    fox
    0.73
     antworte
    0.73
     خوردن
    0.71
     jeweils
    0.71
    Activations Density 0.071%

    No Known Activations