© 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. 10-GEMMASCOPE-RES-131K
    7. 127214
    Prev
    Next
    INDEX
    Explanations

    before events or states

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

    This neuron fires strongly on short 3-word spans of content tokens (trigrams).

    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
     când
    -1.55
     not
    -1.52
     összefoglal
    -1.49
     princípio
    -1.38
    少々
    -1.38
     работают
    -1.38
     biraz
    -1.37
    仕事を
    -1.34
     bijna
    -1.34
    苙
    -1.34
    POSITIVE LOGITS
     or
    1.68
     we
    1.66
     menangis
    1.30
     årets
    1.29
     any
    1.29
     или
    1.29
     our
    1.27
     Selasa
    1.26
    ,”
    1.23
     sondern
    1.22
    Activations Density 0.156%

    No Known Activations