© 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. 39071
    Prev
    Next
    INDEX
    Explanations

    did followed by verb

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

    The neuron detects occurrences of the past‐tense auxiliary “did.”

    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
     is
    -2.38
    ,
    -2.09
     frankly
    -2.00
    e
    -1.98
     other
    -1.95
     similar
    -1.95
    0
    -1.93
    ↵↵
    -1.88
     how
    -1.84
    A
    -1.74
    POSITIVE LOGITS
    ixante
    1.65
    轳
    1.59
     csrf
    1.59
     něko
    1.56
     motivos
    1.55
    🏃
    1.52
    😺
    1.52
     ;"
    1.52
    悝
    1.52
    ormais
    1.50
    Activations Density 0.009%

    No Known Activations