© 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. 85987
    Prev
    Next
    INDEX
    Explanations

    certain conditions or alternatives

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

    This neuron predominantly activates on common English function words (e.g. determiners, prepositions, and relative pronouns like “the,” “a,” “to,” “as,” “which,” “also”).

    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
     chaus
    -1.59
     durer
    -1.58
     procéder
    -1.57
     diminuer
    -1.56
     alho
    -1.56
     épais
    -1.55
     appeler
    -1.51
     poches
    -1.51
     corriger
    -1.50
     terminer
    -1.50
    POSITIVE LOGITS
     or
    2.38
     and
    1.99
     which
    1.84
     we
    1.73
     how
    1.68
     so
    1.56
     where
    1.45
     than
    1.41
     because
    1.39
     it
    1.36
    Activations Density 0.368%

    No Known Activations