© 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. 72291
    Prev
    Next
    INDEX
    Explanations

    subject and predicate

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

    The neuron predominantly activates on numeric tokens or numeric sequences (i.e. digits and numbers) in the text.

    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
     merupakan
    -1.04
     kritik
    -1.00
     kapak
    -1.00
     klinik
    -0.98
     écou
    -0.97
     durant
    -0.96
    ARU
    -0.95
     hiperten
    -0.94
     taraf
    -0.94
    Spesifikasi
    -0.93
    POSITIVE LOGITS
     based
    0.92
    cyan
    0.92
     Sünde
    0.90
    窺
    0.88
     what
    0.88
    搡
    0.85
     Росії
    0.84
     accompanied
    0.83
     there
    0.82
     requirements
    0.81
    Activations Density 0.050%

    No Known Activations