© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 62776
    Prev
    Next
    INDEX
    Explanations

    grammar and word choice

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

    This neuron spots the individual word or phrase options being compared (the quoted or listed candidate forms) in usage‐and‐grammar questions.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    Tidak
    -1.17
     tristes
    -1.12
    ...".
    -1.06
    ...',
    -1.05
    ffichage
    -1.03
    Τι
    -0.98
     فقط
    -0.98
     aiment
    -0.96
    ...",
    -0.96
    Đối
    -0.95
    POSITIVE LOGITS
     which
    1.16
     since
    1.16
     that
    1.13
     until
    1.13
     had
    1.09
     nor
    1.02
     being
    1.01
     for
    1.00
     was
    0.98
     have
    0.98
    Activations Density 0.005%

    No Known Activations