© 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. 22-GEMMASCOPE-RES-131K
    4. 98685
    Prev
    Next
    INDEX
    Explanations

    code and text interfaces

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

    The neuron consistently lights up on code‐structure delimiters—especially closing braces, “endif”/block‐end markers, semicolons, and other tokens that terminate statements or blocks.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
    Condol
    -0.95
    interess
    -0.92
    haviours
    -0.91
     couteau
    -0.88
    Tämä
    -0.88
    Usual
    -0.88
     nocturna
    -0.87
     作り方
    -0.87
     próximo
    -0.85
    汙
    -0.85
    POSITIVE LOGITS
    </b>
    0.91
     />
    0.90
    <i>
    0.86
    <
    0.85
    ,​
    0.84
     ニュース
    0.84
    NED
    0.83
    </h2>
    0.83
     Ñ
    0.83
    frutar
    0.82
    Activations Density 0.004%

    No Known Activations