© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 42297
    Prev
    Next
    INDEX
    Explanations

    partially word fragments

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

    The neuron fires strongly on subword pieces that begin a word (i.e. the first token after a space).

    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
     which
    -2.50
     with
    -2.41
     when
    -2.31
     in
    -2.30
     while
    -2.28
     where
    -2.27
     cannot
    -2.19
     all
    -2.17
     of
    -2.16
     but
    -2.09
    POSITIVE LOGITS
     pistolet
    2.08
     plomb
    1.91
     remé
    1.86
     ۴
    1.85
     camiset
    1.81
     própri
    1.79
     anún
    1.78
    쨩
    1.78
    놋
    1.78
     avocat
    1.77
    Activations Density 0.062%

    No Known Activations