© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 85337
    Prev
    Next
    INDEX
    Explanations

    didn't + verb

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

    The neuron fires on the terminal “t” in negative contractions (e.g. don’t, didn’t, couldn’t).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     are
    -3.06
     尔
    -2.72
    횃
    -2.67
    jeeling
    -2.63
    呖
    -2.61
    ⚋
    -2.59
    érieure
    -2.47
     Bedürfn
    -2.45
    樒
    -2.45
    mathrm
    -2.44
    POSITIVE LOGITS
    t
    3.31
    ли
    2.61
    ↵
    2.48
    </h3>
    2.47
    $
    2.39
     XVII
    2.16
    Doesn
    2.11
    「
    2.09
     ‘’
    2.08
    le
    2.06
    Activations Density 0.092%

    No Known Activations