© 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. 24-GEMMASCOPE-2-RES-16K
    7. 14614
    Prev
    Next
    INDEX
    Explanations

    preposition object structure

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

    The neuron is essentially flagging uncommon or low‐frequency content words (i.e. rarer tokens) in the text.

    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_24_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
     Schuld
    0.82
    ἲ
    0.80
    學者
    0.75
    ⣀
    0.75
     учены
    0.75
     сообщи
    0.74
     حسب
    0.74
    vý
    0.73
    ammlung
    0.72
    {-\
    0.71
    POSITIVE LOGITS
    EN
    0.80
    يه
    0.73
    数字化
    0.72
    вається
    0.70
    TER
    0.69
    Aph
    0.69
     numerosi
    0.69
     タイプ
    0.68
    urée
    0.67
    が多い
    0.66
    Activations Density 0.000%

    No Known Activations