© 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. 33535
    Prev
    Next
    INDEX
    Explanations

    mathematical and programming expressions

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

    The neuron detects the occurrence of the word “for” (as in “…=… for x”) in math problem statements.

    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
     norvég
    -1.17
    красный
    -1.12
     incess
    -1.01
    recuper
    -1.01
     comanda
    -1.00
    ⇲
    -0.97
     ortodo
    -0.97
    Reduces
    -0.96
     comprimidos
    -0.95
     xadrez
    -0.95
    POSITIVE LOGITS
     also
    1.27
    ;
    0.98
    L
    0.93
    ::
    0.93
    M
    0.92
     веке
    0.92
     C
    0.90
    C
    0.90
    R
    0.90
    ж
    0.89
    Activations Density 0.033%

    No Known Activations