© 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. 22-GEMMASCOPE-RES-131K
    7. 33888
    Prev
    Next
    INDEX
    Explanations

    past perfect tense

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

    The neuron reliably activates on occurrences of the auxiliary “had,” marking that specific past-tense verb form.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     that
    -1.98
     and
    -1.87
    ៈ
    -1.74
    anggap
    -1.63
    aritions
    -1.61
    PENDAHULUAN
    -1.60
     bahkan
    -1.59
    ﹔
    -1.51
     asegurado
    -1.51
    ťaž
    -1.51
    POSITIVE LOGITS
     been
    2.36
     "
    1.82
     not
    1.63
    м
    1.63
    ता
    1.62
     become
    1.45
    ’
    1.41
    He
    1.40
     now
    1.39
    薔薇
    1.37
    Activations Density 0.025%

    No Known Activations