© 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. 2267
    Prev
    Next
    INDEX
    Explanations

    separate

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

    The neuron activates on words expressing the idea of being “separate” or “separation,” e.g. separation, separate, independently, separately.

    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
    нома
    0.57
    苃
    0.54
    អេ
    0.51
    蹉
    0.51
    ewód
    0.50
    ༦
    0.49
    ⏟
    0.49
    応用
    0.49
    WithObjects
    0.49
     begeistert
    0.48
    POSITIVE LOGITS
     separate
    4.03
    separate
    3.78
     seperate
    3.77
     Separate
    3.72
    Separate
    3.67
     separately
    3.55
     separ
    3.38
     separado
    3.27
     отде
    3.23
     পৃথক
    3.17
    Activations Density 1.028%

    No Known Activations