© 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. 71594
    Prev
    Next
    INDEX
    Explanations

    nan + word part

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

    This neuron responds to any token containing the substring “nan” (whether in place names like “Nan…” or in “nano-” prefixes such as “nanotubes”).

    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
    ：“
    -3.00
    ，“
    -2.84
    焌
    -2.81
     beeindruck
    -2.69
     unglaub
    -2.69
    、“
    -2.63
    鉛筆
    -2.48
    
    -2.44
     verbess
    -2.31
    一種
    -2.31
    POSITIVE LOGITS
        
    2.78
     «
    2.70
     возможно
    2.61
    و
    2.59
    </h2>
    2.53
    `
    2.47
    u
    2.44
    ?
    2.38
    我们就
    2.36
    H
    2.36
    Activations Density 0.016%

    No Known Activations