© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 15921
    Prev
    Next
    INDEX
    Explanations

    print and class definitions

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

    The neuron fires most strongly on long runs of leading whitespace (indentation) in code or formatted text.

    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
    了什么
    -1.20
     Bizarre
    -1.09
     vormen
    -1.09
     möglicherweise
    -1.08
     Reglamento
    -1.06
    可能會
    -1.05
     blossoming
    -1.04
    Zeneca
    -1.04
     mago
    -1.02
     devising
    -1.02
    POSITIVE LOGITS
    .“
    1.08
     geen
    1.06
     brune
    1.04
     stanovnika
    1.03
     if
    1.03
     hadde
    1.01
    "],
    1.01
     hæ
    0.97
     –
    0.96
     für
    0.96
    Activations Density 0.029%

    No Known Activations