© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 26850
    Prev
    Next
    INDEX
    Explanations

    URLs and specific entities

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

    The neuron is looking for URL‐ or HTML‐encoding tokens—i.e. percent–encoded sequences (like “%3A”, “%2F”) and numeric character references (like “/”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     that
    -1.48
     other
    -1.38
     having
    -1.36
     enhancing
    -1.34
     carnaval
    -1.30
     provide
    -1.30
     even
    -1.27
     have
    -1.26
     both
    -1.24
     taking
    -1.23
    POSITIVE LOGITS
     impresionantes
    1.48
     colossal
    1.36
     revamped
    1.35
     unending
    1.34
    horloge
    1.32
     opulent
    1.30
     essayé
    1.27
     marvelous
    1.25
     gigantic
    1.23
     monstrous
    1.21
    Activations Density 0.004%

    No Known Activations