© 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. 10-GEMMASCOPE-RES-131K
    7. 99896
    Prev
    Next
    INDEX
    Explanations

    casting spells

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

    The neuron fires whenever it sees the word “spell” (in any casing or context).

    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
    馀
    -2.52
    䊳
    -2.50
    絰
    -2.34
    猛然
    -2.33
     passe
    -2.30
    这本书
    -2.30
     he
    -2.28
    ఽ
    -2.25
    
    -2.19
    —-
    -2.17
    POSITIVE LOGITS
    “
    3.11
     If
    2.77
    可以
    2.61
     rimasto
    2.56
    月
    2.53
    也
    2.53
    S
    2.48
    a
    2.42
    just
    2.36
     він
    2.36
    Activations Density 0.005%

    No Known Activations