© 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. 3071
    Prev
    Next
    INDEX
    Explanations

    should be, never, not

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

    This neuron strongly responds to the modal auxiliary “should.”

    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
     wounding
    0.69
     uncharted
    0.67
     necessità
    0.67
     wounded
    0.66
     factions
    0.66
    <td>
    0.65
     ממ
    0.64
     opportune
    0.63
    য়োজনীয়
    0.63
     elusive
    0.63
    POSITIVE LOGITS
    Should
    1.32
     ideally
    1.25
     debería
    1.20
     dovrebbe
    1.20
     powinno
    1.20
    should
    1.19
    Ideally
    1.19
     powinien
    1.19
     Shouldn
    1.18
    应
    1.16
    Activations Density 0.183%

    No Known Activations