© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 60422
    Prev
    Next
    INDEX
    Explanations

    writing, inscription, lettering

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

    The neuron is detecting numeric tokens—digits and number sequences (e.g. measurements, years, counts).

    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
    Пред
    -0.91
    ajā
    -0.84
    DISCUSSION
    -0.83
    ibase
    -0.82
    از
    -0.78
     noastră
    -0.78
     vklju
    -0.77
    hibli
    -0.77
    Россия
    -0.77
    signable
    -0.77
    POSITIVE LOGITS
     words
    3.19
     writing
    2.48
     inscription
    2.47
     text
    2.39
     writings
    2.34
     wording
    2.27
     lettering
    2.23
    文字
    2.22
     word
    2.16
     inscriptions
    2.14
    Activations Density 0.103%

    No Known Activations