© 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. 36449
    Prev
    Next
    INDEX
    Explanations

    code definition and scripting

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

    The neuron consistently fires on numeric tokens (e.g. integer literals) in code or data.

    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
     Each
    -1.09
    Those
    -1.00
    ｪ
    -0.99
     estrenar
    -0.98
    houettes
    -0.98
     Gastgeber
    -0.98
    alamu
    -0.90
    ทีม
    -0.90
     It
    -0.90
    🦦
    -0.90
    POSITIVE LOGITS
     Gundam
    1.14
     sic
    1.02
     uner
    1.00
    Hdr
    0.97
     cristian
    0.96
    Hoog
    0.88
     cae
    0.88
     nei
    0.87
    Imgs
    0.87
    perhaps
    0.87
    Activations Density 0.004%

    No Known Activations