© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 77276
    Prev
    Next
    INDEX
    Explanations

    what is happening or wrong

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

    The neuron strongly activates on the question word “what.”

    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
     Specifies
    -1.03
     Allows
    -1.02
     Ahnung
    -1.02
    tembro
    -1.00
    quelize
    -1.00
    >=</
    -0.99
     vélos
    -0.99
    儛
    -0.98
     autorité
    -0.98
    めちゃ
    -0.97
    POSITIVE LOGITS
     to
    2.52
     kind
    1.97
     else
    1.66
     is
    1.65
     the
    1.61
     type
    1.59
     might
    1.56
     we
    1.55
     happened
    1.52
     it
    1.48
    Activations Density 0.046%

    No Known Activations