© 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. 71022
    Prev
    Next
    INDEX
    Explanations

    programming code constructs

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

    The neuron activates on programming‐language tokens—especially identifiers (like variable names) and adjoining punctuation (e.g. brackets)—marking bits of source code rather than natural language.

    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
    vature
    -0.91
    titanic
    -0.89
    着她
    -0.87
    meiras
    -0.85
    cửa
    -0.85
    んじゃ
    -0.82
    shred
    -0.82
     curieux
    -0.82
     vulgaires
    -0.81
     diperhatikan
    -0.80
    POSITIVE LOGITS
     between
    0.98
    ഐ
    0.90
     called
    0.89
     proprietors
    0.88
     that
    0.85
     allora
    0.84
     enige
    0.82
     fuckin
    0.81
     entweder
    0.81
    ebbe
    0.80
    Activations Density 0.009%

    No Known Activations