© 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. 47590
    Prev
    Next
    INDEX
    Explanations

    scaffold, scaffolding, form work

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

    The neuron is highly sensitive to formal technical or patent–style language (e.g., descriptive mechanical terms and structure‐related words).

    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
     BORDER
    -0.79
     Betty
    -0.79
     beep
    -0.78
    estream
    -0.77
    Dyna
    -0.75
    Betty
    -0.75
    TokenName
    -0.75
    цепт
    -0.73
     Elsa
    -0.72
    unsqueeze
    -0.70
    POSITIVE LOGITS
     scaffold
    1.25
    Scaffold
    1.16
     Scaffold
    1.13
    scaffold
    1.08
     scaffolding
    1.04
     scaffolds
    1.01
     huur
    0.95
     safety
    0.94
     platforms
    0.93
     worker
    0.88
    Activations Density 0.009%

    No Known Activations