© 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. 26719
    Prev
    Next
    INDEX
    Explanations

    mathematical fractions and divisions

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

    The neuron is looking for standalone numeric tokens (numbers and numeric constants) in mathematical expressions.

    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
    為に
    -1.52
     like
    -1.22
    に入れる
    -1.20
     ilustracji
    -1.19
    様な
    -1.16
    ところに
    -1.16
     will
    -1.09
     it
    -1.09
     твой
    -1.09
     something
    -1.06
    POSITIVE LOGITS
    そして
    1.53
    <bos>
    1.47
    がありますが
    1.46
    があり
    1.41
    ですが
    1.35
    と
    1.34
    تحميل
    1.34
     ObjectParameter
    1.32
     CascadeType
    1.29
    は
    1.28
    Activations Density 0.023%

    No Known Activations