© 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. 48374
    Prev
    Next
    INDEX
    Explanations

    think twice

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

    The neuron activates on the word “think” (and its inflected forms like “thinks,” “thought,” “thinking”).

    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
    throwable
    -0.81
    TableView
    -0.80
    плення
    -0.78
    gique
    -0.77
    filters
    -0.74
    erv
    -0.72
    unitOfWork
    -0.72
    ȫ
    -0.70
     znaj
    -0.70
     of
    -0.69
    POSITIVE LOGITS
     hard
    3.34
     harder
    2.38
    hard
    2.34
    Hard
    2.28
     Hard
    2.11
    HARD
    1.95
     hardest
    1.86
     long
    1.82
     carefully
    1.73
     HARD
    1.73
    Activations Density 0.021%

    No Known Activations