© 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. 6424
    Prev
    Next
    INDEX
    Explanations

    between items

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

    The neuron strongly activates on English negative contractions ending in “n’t” (e.g. couldn’t, wasn’t, didn’t).

    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
    ੳ
    0.68
    udson
    0.58
    цов
    0.58
    infty
    0.58
    ду
    0.57
    iftoire
    0.56
     đc
    0.56
    uyền
    0.55
    де
    0.55
     মোঃ
    0.55
    POSITIVE LOGITS
     bandages
    0.60
     resorting
    0.59
    ​—
    0.58
     bystanders
    0.57
     unbek
    0.56
     enacting
    0.55
     conversaciones
    0.55
     beneath
    0.55
     Aunt
    0.54
     beside
    0.54
    Activations Density 0.133%

    No Known Activations