© 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. 34-GEMMASCOPE-RES-131K
    4. 1938
    Prev
    Next
    INDEX
    Explanations

    close, dispose, hide actions

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

    The neuron strongly activates on programming‐language punctuation and structural tokens—braces, parentheses, indentation spaces, arrows, etc.—i.e. the syntax elements of 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_34/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
    вр
    -0.82
     Fehler
    -0.79
    čného
    -0.73
     ફ
    -0.71
    Alerter
    -0.71
    🅡
    -0.71
     Arzneimit
    -0.71
    ilistic
    -0.69
     Baus
    -0.69
     sæ
    -0.69
    POSITIVE LOGITS
     this
    1.89
    dispose
    1.68
    setVisible
    1.66
    this
    1.60
     dispose
    1.34
     visible
    1.23
    Close
    1.22
    Hide
    1.18
    Visible
    1.15
    Dispose
    1.05
    Activations Density 0.005%

    No Known Activations