© 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. 48022
    Prev
    Next
    INDEX
    Explanations

    status flags and control registers

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

    The neuron spikes on uppercase register‐style labels and section headers in low‐level code (e.g. “Low Voltage Detect Status and Control 1 Register”), effectively detecting hardware register names in comments or definitions.

    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
     Desmond
    -0.88
     similaire
    -0.80
    utebol
    -0.79
     Bere
    -0.74
     Baden
    -0.74
    콧
    -0.71
    орк
    -0.71
    Baden
    -0.71
     Keegan
    -0.68
    ὸ
    -0.68
    POSITIVE LOGITS
     cleared
    1.20
    Cleared
    1.12
    cleared
    1.02
     flag
    0.96
     status
    0.86
     cause
    0.86
     sticky
    0.86
     pending
    0.85
     Masked
    0.85
     flags
    0.85
    Activations Density 0.015%

    No Known Activations