© 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. 22-GEMMASCOPE-RES-131K
    7. 79890
    Prev
    Next
    INDEX
    Explanations

    code structure tokens

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

    The neuron strongly activates on code-declaration markers (e.g. “type,” “#define,” “#pragma,” attribute brackets), effectively spotting language keywords and preprocessor directives that introduce or define constructs.

    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
     of
    -1.21
     for
    -1.19
     (
    -1.11
    VALIDATE
    -1.06
    ják
    -1.02
    ;
    -0.98
     urbano
    -0.98
    вец
    -0.98
     як
    -0.98
    лайн
    -0.98
    POSITIVE LOGITS
    푉
    1.44
    Rote
    1.41
    Missi
    1.38
    GDPR
    1.38
     pubblicità
    1.31
     volontà
    1.30
     piacevole
    1.28
    Nero
    1.27
    Mosa
    1.27
    Steg
    1.24
    Activations Density 0.000%

    No Known Activations