© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 13792
    Prev
    Next
    INDEX
    Explanations

    mathematical expressions with variables

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

    The neuron fires on isolated one‐character symbols—especially single‐letter variables (e.g. d, x, y, z, u, t) and single‐digit numbers—i.e. the atomic tokens used in math or symbolic 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_10/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
     it
    -1.57
     of
    -1.42
     asust
    -1.41
     tonel
    -1.39
     pág
    -1.38
     plátano
    -1.36
     its
    -1.34
     saber
    -1.30
     harina
    -1.30
     tujuh
    -1.30
    POSITIVE LOGITS
     that
    1.36
     andreas
    1.30
     facendo
    1.30
     pablo
    1.27
    f
    1.26
     clipart
    1.26
     که
    1.24
     töö
    1.23
    op
    1.23
    in
    1.23
    Activations Density 0.107%

    No Known Activations