© 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. 76674
    Prev
    Next
    INDEX
    Explanations

    separator before a new statement

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

    This neuron fires on the em-dash separators used in news‐style datelines (e.g. “NEW YORK—” or “WASHINGTON—”).

    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
     there
    -1.20
    hear
    -0.96
     we
    -0.95
     solicitó
    -0.93
    podob
    -0.92
     вопросов
    -0.92
    ñó
    -0.90
    开口道
    -0.90
    ijų
    -0.90
    吭
    -0.89
    POSITIVE LOGITS
     noting
    2.73
     citing
    2.45
     referring
    2.41
     pointing
    2.36
     referencing
    1.84
    refer
    1.75
    pointing
    1.73
     stating
    1.66
     calling
    1.65
    Point
    1.58
    Activations Density 0.057%

    No Known Activations