© 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. 86813
    Prev
    Next
    INDEX
    Explanations

    early hours/morning/night

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

    The neuron flags temporal expressions—dates, days of the week, times, and parts of day (e.g. “early Saturday morning,” “late Monday night,” month names, clock times).

    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
    下午
    -1.34
    Lunch
    -0.92
     Lunch
    -0.90
    allon
    -0.82
     lunch
    -0.80
    晚上
    -0.77
    Afternoon
    -0.76
    中午
    -0.73
    йлер
    -0.73
     вечером
    -0.73
    POSITIVE LOGITS
     madrugada
    2.94
     early
    2.91
    early
    2.33
    Early
    2.27
     Early
    2.09
    EARLY
    1.87
     dawn
    1.84
     EARLY
    1.75
     morning
    1.70
     overnight
    1.63
    Activations Density 0.047%

    No Known Activations