© 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. 93802
    Prev
    Next
    INDEX
    Explanations

    names following titles/descriptions

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

    The neuron fires on tokens that are capitalized—especially proper names, section‐header words, or the first word of sentences.

    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
     apartament
    -1.45
     revisor
    -1.33
     conden
    -1.23
    例文
    -1.21
     bantu
    -1.20
     lehen
    -1.20
     susten
    -1.19
     sigu
    -1.19
     elet
    -1.18
     Svenska
    -1.17
    POSITIVE LOGITS
     proporcionan
    1.36
     imprimer
    1.34
    ,"
    1.33
     proporciona
    1.32
     esfuer
    1.25
    ⅞
    1.24
    ⅝
    1.23
     nervios
    1.22
     vientos
    1.22
     incluyen
    1.22
    Activations Density 0.240%

    No Known Activations