© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 103868
    Prev
    Next
    INDEX
    Explanations

    withdrawal and retreat

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

    The neuron detects words and forms that signify pulling back or retreating (e.g., “retreat,” “withdraw,” “back off,” “withdrawal”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     sobre
    -1.70
    ве
    -1.60
    </strong>
    -1.55
    .*")]
    -1.52
     cuando
    -1.52
     preocupaciones
    -1.49
     otros
    -1.48
     男人
    -1.43
     reason
    -1.38
     love
    -1.38
    POSITIVE LOGITS
    c
    1.66
     comprenant
    1.66
    igns
    1.59
    🤺
    1.57
    ́n
    1.55
    sierung
    1.52
    Уважа
    1.52
    Высота
    1.50
     pelaku
    1.49
     apoder
    1.46
    Activations Density 0.025%

    No Known Activations