© 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. 73835
    Prev
    Next
    INDEX
    Explanations

    early morning times

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find numeric time‐stamp tokens (digits, colons, “a.m.”/“p.m.” markers) in the text.

    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
     evening
    -1.09
    8
    -0.91
     afternoon
    -0.88
    下午
    -0.84
     wiecz
    -0.82
     вечером
    -0.81
     noon
    -0.79
    Afternoon
    -0.78
     помі
    -0.76
     eighties
    -0.76
    POSITIVE LOGITS
     madrugada
    1.16
    凌晨
    1.14
    午前
    0.85
     AM
    0.79
    kwür
    0.75
    三四
    0.75
    am
    0.75
     kde
    0.74
    AM
    0.73
    обходимо
    0.73
    Activations Density 0.025%

    No Known Activations