© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 68100
    Prev
    Next
    INDEX
    Explanations

    higher or lower states

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

    This neuron fires on comparative terms indicating a reduction—most notably the word “lower.”

    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
    </b>
    -2.47
    穑
    -2.33
     The
    -2.25
    zości
    -2.25
    خصة
    -2.23
    ⽬
    -2.16
    ebab
    -2.16
     This
    -2.14
    赝
    -2.14
    来不及
    -2.11
    POSITIVE LOGITS
     than
    4.19
    </h3>
    3.77
    most
    2.92
    $
    2.80
    7
    2.70
    that
    2.69
    3
    2.64
    what
    2.61
    ان
    2.45
    of
    2.42
    Activations Density 0.034%

    No Known Activations