© 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. 24-GEMMASCOPE-2-RES-16K
    7. 3728
    Prev
    Next
    INDEX
    Explanations

    can and cannot

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

    The neuron predominantly activates on occurrences of normative modal verbs and negation (e.g. “should,” “shouldn’t,” “not”) indicating advisory or prescriptive language.

    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
    وني
    0.74
    ന്ത
    0.74
     যাইতে
    0.72
    л
    0.71
    وان
    0.69
    وس
    0.68
     กี้
    0.68
    ໃຈ
    0.68
    дня
    0.67
    ទទ
    0.66
    POSITIVE LOGITS
     melakukannya
    0.83
    以及
    0.79
    ;
    0.76
    함으로써
    0.75
     以及
    0.73
     ซึ่ง
    0.71
     由于
    0.71
     ovvero
    0.71
    ).
    0.71
     bởi
    0.70
    Activations Density 0.070%

    No Known Activations