© 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. 16370
    Prev
    Next
    INDEX
    Explanations

    sequence after specified words

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

    The neuron fires on relational connectives—common function words (especially prepositions like “of,” “with,” “into,” “respect,” etc.) that link or relate parts of a phrase.

    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
    шивка
    -0.94
    ciun
    -0.90
     gobierno
    -0.86
     pleno
    -0.85
     тера
    -0.85
     keja
    -0.84
    왠
    -0.84
     phú
    -0.83
     polymorphisms
    -0.82
    られて
    -0.82
    POSITIVE LOGITS
     seules
    0.93
    ülen
    0.93
     ſever
    0.90
    JoinColumn
    0.89
    ⁰
    0.89
    providedIn
    0.87
     unſ
    0.87
    Kapcsolódó
    0.87
    toupper
    0.85
    мимо
    0.85
    Activations Density 0.046%

    No Known Activations