© 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. 34-GEMMASCOPE-RES-131K
    7. 31483
    Prev
    Next
    INDEX
    Explanations

    divine grace and mercy

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

    The neuron fires strongly on explicit references to salvation by grace (e.g. the “saved-by-grace” phrasing and related theological keywords like grace, love, truth).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Interpre
    -0.90
     Tactical
    -0.77
    MSS
    -0.72
     Rink
    -0.70
    ibel
    -0.70
     Villarreal
    -0.70
     Katrina
    -0.70
     nau
    -0.69
    inkler
    -0.69
     DED
    -0.69
    POSITIVE LOGITS
     grace
    4.22
     Grace
    3.48
    grace
    3.23
    Grace
    3.23
     GRACE
    2.66
    GRACE
    2.38
     graça
    1.79
     gracia
    1.74
     Gnade
    1.72
     mercy
    1.71
    Activations Density 0.023%

    No Known Activations