© 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. 22-GEMMASCOPE-RES-131K
    4. 54542
    Prev
    Next
    INDEX
    Explanations

    code snippets or structures

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

    The neuron activates on programming‐language tokens (keywords, identifiers, numbers, punctuation), i.e. pieces of source code rather than natural‐language 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
     mesaj
    -0.98
    ums
    -0.98
    FirstResponder
    -0.98
    口中
    -0.98
    upuesto
    -0.97
    ситет
    -0.95
    ondissement
    -0.95
    äldrar
    -0.94
     салат
    -0.94
    魄
    -0.94
    POSITIVE LOGITS
     two
    1.70
     three
    1.22
     both
    1.18
    Two
    1.15
     but
    1.12
     only
    1.11
     wynosi
    1.10
     каждый
    1.10
     Both
    1.09
     Каждый
    1.09
    Activations Density 0.196%

    No Known Activations