© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 9220
    Prev
    Next
    INDEX
    Explanations

    consider / take into account

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

    The neuron fires on user‐experience or interface‐design terminology—words about viewers/users, designers, and keeping the user’s mind or experience in focus.

    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
     Mont
    0.55
     ਸਾ
    0.52
     hailed
    0.51
    itzar
    0.50
    〈
    0.50
     héros
    0.49
     كده
    0.49
    rég
    0.49
     quickest
    0.49
    éndose
    0.49
    POSITIVE LOGITS
    考慮
    2.58
     consideration
    2.45
     considerations
    2.39
     Consideration
    2.36
     berücksichtigt
    2.36
     berücksicht
    2.30
    考虑
    2.28
     고려
    2.27
     Considerations
    2.23
     учитывать
    2.09
    Activations Density 0.425%

    No Known Activations