© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 10932
    Prev
    Next
    INDEX
    Explanations

    like a/the, if searching

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

    The neuron detects comparative simile constructions signaled by words like “like” or “as” (e.g. “bathing her is like bathing a piglet,” “deaf as a post,” “turning to ice”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     apparently
    0.64
    apparently
    0.63
    like
    0.62
     तरह
    0.61
     like
    0.60
     noises
    0.59
     rhymes
    0.59
     feeling
    0.58
     signs
    0.57
     okaza
    0.55
    POSITIVE LOGITS
     sandpaper
    0.68
     molasses
    0.66
    洪水
    0.65
    小学生
    0.64
     heaven
    0.60
     slowmotion
    0.60
     বন্যার
    0.59
    天堂
    0.57
     fueran
    0.57
     maharaja
    0.57
    Activations Density 0.313%

    No Known Activations