© 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. 40038
    Prev
    Next
    INDEX
    Explanations

    code identifiers

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

    The neuron fires on numeric literals (e.g. integers) embedded in code.

    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
    ()");
    -1.08
    ());
    
    -1.05
    於
    -0.93
     bekannt
    -0.91
    </h2>
    -0.91
     Despite
    -0.91
     whoſe
    -0.91
    ());
    -0.91
    '];
    
    -0.88
    ()">
    -0.88
    POSITIVE LOGITS
    presumably
    0.89
     anything
    0.88
     such
    0.85
     dore
    0.85
     others
    0.84
     VIET
    0.83
    seems
    0.82
    actually
    0.82
     estadía
    0.81
    を重ね
    0.81
    Activations Density 0.040%

    No Known Activations