© 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. 128763
    Prev
    Next
    INDEX
    Explanations

    The neuron activates on words that introduce or label numerical data or measurements (e.g. “ratio,” “values,” “code,” “sugar”)—i.e. tokens that signal quantitative information.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    of [UNIT]

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    ？』
    -1.14
    zyści
    -1.07
     other
    -1.04
     not
    -0.98
    Fourni
    -0.98
     substantially
    -0.98
     plethora
    -0.95
    などを
    -0.94
    Cabbage
    -0.94
     superintend
    -0.93
    POSITIVE LOGITS
    if
    1.18
     België
    1.16
     dziś
    1.12
     pueden
    1.11
    middels
    1.04
    alas
    1.02
     you
    1.02
    corrhi
    1.02
     podrás
    1.01
    ocardial
    1.01
    Activations Density 0.071%

    No Known Activations