© 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. 10-GEMMASCOPE-RES-131K
    7. 44448
    Prev
    Next
    INDEX
    Explanations

    description

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

    The neuron is tuned to spot numeric tokens and simple numeric expressions (digits, multi‐digit numbers, hyphenated numbers, and related symbols).

    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
     tromper
    -1.63
     diriger
    -1.54
    quese
    -1.43
     véritables
    -1.41
     énorm
    -1.40
     raconter
    -1.40
     quieran
    -1.39
     októ
    -1.39
     nieuwe
    -1.34
    耷
    -1.34
    POSITIVE LOGITS
     of
    2.39
     we
    1.80
     it
    1.80
     including
    1.70
     before
    1.64
     my
    1.61
     at
    1.55
     our
    1.55
     such
    1.52
     an
    1.43
    Activations Density 0.026%

    No Known Activations