© 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. 12773
    Prev
    Next
    INDEX
    Explanations

    system command execution

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

    The neuron detects object-member access using the dot operator (e.g. “object.member” or “response.Output.Write”).

    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
     garagem
    -1.02
    hender
    -0.93
    ஸ
    -0.93
    ouncements
    -0.93
     damskie
    -0.91
     antigo
    -0.89
     przód
    -0.89
    화
    -0.88
     пръ
    -0.87
    hende
    -0.87
    POSITIVE LOGITS
     despite
    0.95
     by
    0.94
    復興
    0.89
     although
    0.86
     variante
    0.84
     标签
    0.84
     corrector
    0.84
     они
    0.84
     when
    0.83
    caccio
    0.83
    Activations Density 0.005%

    No Known Activations