© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 91885
    Prev
    Next
    INDEX
    Explanations

    extending horizontally or vertically

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

    This neuron specializes in spotting formal patent‐style specification language—i.e. component names, enumerated parts, and structural relationship terms in technical descriptions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Produkten
    -0.94
    じゃなくて
    -0.89
     kuan
    -0.87
     tutur
    -0.86
     mengucap
    -0.85
    ácie
    -0.85
     Agregar
    -0.83
    Página
    -0.82
    alisco
    -0.82
     ortodo
    -0.81
    POSITIVE LOGITS
     its
    1.02
     ends
    0.99
     inner
    0.89
     boss
    0.88
     противополо
    0.86
    micamente
    0.85
    形成
    0.85
     bosses
    0.83
     opposite
    0.82
    inner
    0.81
    Activations Density 0.793%

    No Known Activations