© 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. 22-GEMMASCOPE-RES-131K
    7. 72000
    Prev
    Next
    INDEX
    Explanations

    properties of mathematical structures

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

    The neuron fires on isolated mathematical-concept keywords (e.g. subgroup, cyclic, diagonalizable, hereditary, etc.).

    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
     amazingly
    -1.32
     insanely
    -1.30
     incredibly
    -1.27
     chamado
    -1.27
     exceedingly
    -1.23
     homicidio
    -1.21
     extremely
    -1.20
     ส์
    -1.20
    
    -1.20
     amazing
    -1.19
    POSITIVE LOGITS
     with
    2.16
     after
    1.45
     if
    1.44
     on
    1.29
     at
    1.29
     despite
    1.25
     before
    1.20
     should
    1.20
     while
    1.19
     reduces
    1.17
    Activations Density 0.034%

    No Known Activations