© 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. 65875
    Prev
    Next
    INDEX
    Explanations

    tests and verification

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

    The neuron is primarily detecting numeric constants (e.g. integers, floats, percentages) in code.

    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
    ®:
    -1.28
     kazak
    -1.22
    】:
    -1.20
     kenzo
    -1.20
     hama
    -1.19
     albic
    -1.18
     versace
    -1.17
     bret
    -1.16
     siena
    -1.16
     fabia
    -1.16
    POSITIVE LOGITS
     or
    1.37
     no
    1.08
     when
    1.06
     //
    1.02
     but
    1.02
     not
    1.02
     variedades
    1.01
     něko
    0.99
     lesiones
    0.97
     from
    0.96
    Activations Density 0.005%

    No Known Activations