© 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. 129577
    Prev
    Next
    INDEX
    Explanations

    This neuron activates on longer (multi-syllable) content words—that is, it picks out unusually long tokens.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    asking for

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     who
    -1.38
     which
    -1.18
    ,
    -1.16
    =$((
    -1.09
     aqueles
    -1.07
    
    -1.06
     such
    -1.05
    simum
    -1.05
    化した
    -1.05
     They
    -1.04
    POSITIVE LOGITS
     perawatan
    1.41
     parezca
    1.26
    Nachteile
    1.21
    . 
    1.19
    你的
    1.17
     riveting
    1.16
     menghindari
    1.15
     publicado
    1.14
     brinde
    1.13
     徒歩
    1.11
    Activations Density 0.041%

    No Known Activations