© 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. 10-GEMMASCOPE-RES-131K
    4. 41875
    Prev
    Next
    INDEX
    Explanations

    withdrawn, revoke, retire

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

    The main thing this neuron does is detect words and forms of “withdraw” (e.g. withdrew, withdrawn, withdraws) indicating a removal or retraction.

    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
     will
    -2.25
     still
    -1.87
     are
    -1.69
     The
    -1.59
     soothing
    -1.56
              
    -1.55
     /
    -1.54
     Toten
    -1.51
     cannot
    -1.47
     \
    -1.43
    POSITIVE LOGITS
    CurtirCurtir
    1.89
    跸
    1.77
     começaram
    1.73
    ieť
    1.67
    ToProps
    1.67
     específicamente
    1.59
    alisation
    1.55
    ggiata
    1.49
    嶮
    1.49
    .");
    1.44
    Activations Density 0.014%

    No Known Activations