© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 117002
    Prev
    Next
    INDEX
    Explanations

    long ago, before, after

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

    The neuron activates on the word “long,” especially when used in temporal phrases (e.g. “long ago,” “long forgotten”).

    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
     monstre
    -1.01
    chastic
    -0.88
    があると
    -0.85
     laik
    -0.83
    entire
    -0.82
    <bos>
    -0.79
    mator
    -0.77
    この時期
    -0.76
    geth
    -0.75
    errorHandler
    -0.75
    POSITIVE LOGITS
     ago
    1.72
     long
    1.41
     давно
    1.38
     before
    1.19
     after
    1.05
    之后
    1.03
    Long
    1.02
     afterward
    1.02
    long
    0.99
     AFTER
    0.99
    Activations Density 0.018%

    No Known Activations