© 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. 22-GEMMASCOPE-RES-131K
    7. 33219
    Prev
    Next
    INDEX
    Explanations

    security camera feed

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

    The neuron flags the key subject or object words that name the main topic being asked about (the central technical term or focus noun/verb of a question).

    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
    츰
    -0.86
     pyta
    -0.84
     vertra
    -0.83
    لغة
    -0.82
     plusieurs
    -0.82
    髕
    -0.81
     eenen
    -0.81
     fter
    -0.81
     essayer
    -0.80
     penser
    -0.80
    POSITIVE LOGITS
     when
    1.02
     nearby
    0.89
     from
    0.88
     at
    0.82
    Ꭽ
    0.80
     already
    0.79
     within
    0.78
     on
    0.77
     existing
    0.74
     adven
    0.74
    Activations Density 0.179%

    No Known Activations