© 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. 737
    Prev
    Next
    INDEX
    Explanations

    conditional start with if

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

    This neuron fires on standalone header‐style tokens—words in all caps or title case that serve as section labels or headings (e.g. “TEST,” “SKIP IF,” “Decided,” “James”).

    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
    園
    -0.75
    铆
    -0.73
     recirc
    -0.71
     Ilya
    -0.71
     SAX
    -0.71
    Brace
    -0.68
     HMM
    -0.68
    耐久
    -0.66
    setOpaque
    -0.66
    plit
    -0.66
    POSITIVE LOGITS
    qp
    0.71
     المكت
    0.70
     crescent
    0.70
    QP
    0.68
    ifor
    0.67
    DNF
    0.65
    TextNode
    0.63
    impin
    0.63
     Crescent
    0.63
     sesuai
    0.61
    Activations Density 0.079%

    No Known Activations