© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 20835
    Prev
    Next
    INDEX
    Explanations

    one of the

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

    The neuron responds most strongly to sentence‐ or clause‐separating punctuation (commas and periods) that mark major phrase boundaries.

    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
    nému
    -1.47
    anglais
    -1.38
     into
    -1.36
     they
    -1.33
    やかに
    -1.30
    onlu
    -1.30
    hoodie
    -1.28
    asının
    -1.27
    fæ
    -1.26
     Перейти
    -1.24
    POSITIVE LOGITS
     both
    1.81
    []
    1.55
    \
    1.41
    both
    1.40
     it
    1.35
    它的
    1.34
     when
    1.34
    чень
    1.30
     conhece
    1.28
     departament
    1.26
    Activations Density 0.198%

    No Known Activations