© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 101565
    Prev
    Next
    INDEX
    Explanations

    mathematical and logical expressions

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

    the word "and" in mathematical or technical contexts, particularly in formal problem statements and proofs.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    This neuron activates on numeric tokens containing a decimal point (e.g. “56.5,” “123.5,” “12.75,” 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_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
    idlertid
    -2.33
     toallas
    -2.22
    卌
    -2.19
    konzert
    -2.09
    嗵
    -2.05
    櫺
    -2.02
    茈
    -1.98
    囖
    -1.96
     calefacción
    -1.96
     klasa
    -1.94
    POSITIVE LOGITS
     with
    2.69
     at
    2.69
     When
    2.66
     in
    2.52
    The
    2.44
    '
    2.42
    :
    2.33
    O
    2.27
     this
    2.25
    4
    2.25
    Activations Density 0.062%

    No Known Activations