© 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. 34-GEMMASCOPE-RES-131K
    4. 76190
    Prev
    Next
    INDEX
    Explanations

    license

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

    The neuron strongly activates on code-structure tokens—especially indentation whitespace and block delimiters like braces—i.e. it detects formatting elements that mark code block structure.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    ımıza
    -0.87
    🤨
    -0.84
    ❓
    -0.83
    dresser
    -0.82
    🐢
    -0.81
     בשנת
    -0.80
    煖
    -0.79
    xious
    -0.79
     EVANS
    -0.79
     frågan
    -0.79
    POSITIVE LOGITS
    most
    0.77
     tratan
    0.75
    hig
    0.71
    also
    0.71
    때
    0.71
    unno
    0.69
    usually
    0.69
    ONYMOUS
    0.68
    only
    0.66
    unità
    0.66
    Activations Density 0.241%

    No Known Activations