© 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. 34-GEMMASCOPE-RES-131K
    7. 48780
    Prev
    Next
    INDEX
    Explanations

    prime and composite number factorization

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

    The neuron is specialized for spotting number tokens in math‐problem contexts (e.g. the numeric literals in “prime factors of …,” “greatest common factor of …,” 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_34/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
    FileDialog
    -0.81
    贱
    -0.76
    rumah
    -0.72
    蛋白
    -0.71
     apps
    -0.71
    户
    -0.71
    pinMode
    -0.70
     biting
    -0.69
     knit
    -0.69
     coales
    -0.69
    POSITIVE LOGITS
     factorization
    1.23
     multiplic
    0.91
     Pharisees
    0.88
     composit
    0.87
    composite
    0.86
     composite
    0.83
    FACTOR
    0.79
     COMPOS
    0.79
     ToString
    0.79
     Composite
    0.79
    Activations Density 0.106%

    No Known Activations