© 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. 93635
    Prev
    Next
    INDEX
    Explanations

    code and device identifiers

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

    The neuron strongly activates on programming-style tokens (identifiers, file paths, code keywords, API names, etc.)—in other words, it flags code fragments.

    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
     Tür
    -0.81
    MADRID
    -0.74
    飢
    -0.72
    版
    -0.71
     minValue
    -0.71
     signaling
    -0.71
     what
    -0.70
    𝚐
    -0.70
    늠
    -0.70
    amines
    -0.69
    POSITIVE LOGITS
     Padres
    0.83
     Enc
    0.78
     Thales
    0.77
    popover
    0.77
     epoxy
    0.76
    NEET
    0.75
     Uniwers
    0.75
     Advis
    0.75
    NCT
    0.75
    Dad
    0.74
    Activations Density 0.005%

    No Known Activations