© 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. 65428
    Prev
    Next
    INDEX
    Explanations

    my friend

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

    The neuron sharply activates on numeric tokens—years, counts, floats, and other digits—essentially flagging number‐laden content.

    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
     Dés
    -1.02
     KAY
    -1.02
     FRAN
    -1.01
    寕
    -0.99
     Ramírez
    -0.99
    junge
    -0.99
    personen
    -0.98
     ſever
    -0.98
    alnya
    -0.98
     cannes
    -0.98
    POSITIVE LOGITS
     what
    1.38
     like
    1.22
     we
    1.16
     but
    1.15
     there
    1.12
     What
    1.03
     does
    1.02
     because
    0.98
     why
    0.98
     instead
    0.98
    Activations Density 0.005%

    No Known Activations