© 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. 66734
    Prev
    Next
    INDEX
    Explanations

    Halle Hallelujah

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

    The neuron is highly activated by any token containing a doubled letter (e.g. “Hallett,” “Hallelujah,” “Minnie,” etc.).

    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
     {
    -2.61
    f
    -2.47
    
    -2.42
     этому
    -2.39
    ai
    -2.38
    z
    -2.38
    I
    -2.28
    de
    -2.25
     Inex
    -2.22
    h
    -2.20
    POSITIVE LOGITS
    ſelf
    2.69
    Ꮌ
    2.42
    0
    2.39
    VERY
    2.38
    屨
    2.25
     It
    2.20
     perchè
    2.19
    çok
    2.17
    ︲
    2.17
     augus
    2.14
    Activations Density 0.013%

    No Known Activations