© 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. 33951
    Prev
    Next
    INDEX
    Explanations

    section or arrangement details

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

    The neuron is triggered by leading indentation (runs of whitespace before code tokens), i.e. it flags lines or segments that are indented 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
     Aufführung
    -1.17
     тях
    -1.11
    Ekster
    -1.09
    Along
    -1.06
     zákaz
    -1.06
    多くの
    -1.05
    Different
    -1.03
    Among
    -1.03
    সূত্র
    -1.03
     behandling
    -0.98
    POSITIVE LOGITS
     another
    1.12
     consists
    0.95
     section
    0.94
     comprises
    0.92
    隠
    0.90
     
    0.88
     information
    0.88
     composé
    0.84
     portion
    0.82
     activity
    0.82
    Activations Density 0.004%

    No Known Activations