© 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. 22-GEMMASCOPE-RES-131K
    7. 27895
    Prev
    Next
    INDEX
    Explanations

    model modelLK, AM, numbers, and words

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

    The neuron strongly responds to numeric tokens and related punctuation (numbers, decimals, commas) typically found in product specs or dimension listings.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     it
    -1.27
    อ่าน
    -1.21
     которое
    -1.20
    mathrm
    -1.16
     conecta
    -1.13
    𝑛
    -1.11
    StringProperty
    -1.11
     Setelah
    -1.10
    🧾
    -1.08
     esportes
    -1.07
    POSITIVE LOGITS
     ($("#
    1.35
     Euch
    1.25
     appétit
    1.24
     toate
    1.23
     fiabilidad
    1.20
     vloer
    1.18
    melidir
    1.15
     funcionalidad
    1.15
     zapato
    1.15
     сахара
    1.13
    Activations Density 0.006%

    No Known Activations