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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 6360
    Prev
    Next
    INDEX
    Explanations

    Im- followed by word parts

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

    The neuron highly activates on tokens that begin with the letters “Im” (typically at the start of words like “Immigration,” “Impressive,” or other “Im-” prefixes).

    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
     অবস্থ
    0.63
    아서
    0.57
    alanine
    0.57
    afs
    0.56
    whole
    0.56
    ◞
    0.55
     সফরে
    0.54
     ပ
    0.54
     AUTHOR
    0.54
    ół
    0.54
    POSITIVE LOGITS
     Imp
    0.99
     IMP
    0.94
     Im
    0.93
     impe
    0.87
     imp
    0.86
     impr
    0.80
    Im
    0.79
    IMP
    0.77
     İm
    0.73
     इंप
    0.72
    Activations Density 0.166%

    No Known Activations