© 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. 34-GEMMASCOPE-RES-131K
    7. 93091
    Prev
    Next
    INDEX
    Explanations

    viewDidLoad

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

    The neuron fires on the delimiters and openings of real code blocks (e.g. braces, class/function declarations, doc‐string markers), i.e. it’s detecting structured source‐code regions.

    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
     вле
    -0.90
    often
    -0.82
    usually
    -0.81
    always
    -0.81
     ENV
    -0.80
    ментом
    -0.80
    abonnement
    -0.79
    stes
    -0.79
     Denk
    -0.78
     ส์
    -0.78
    POSITIVE LOGITS
     even
    0.79
    茹
    0.78
    stellung
    0.75
     zobaczyć
    0.75
     cbd
    0.75
     iglesia
    0.75
     Immunol
    0.73
     Ye
    0.73
    yethyl
    0.71
     considers
    0.71
    Activations Density 0.007%

    No Known Activations