© 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. 67652
    Prev
    Next
    INDEX
    Explanations

    new or different concepts

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

    The neuron fires on hyperbolic novelty and emphasis words—tokens like “whole,” “new,” “distinct,” “different,” “level,” or “way” that signal “a whole new/ different level/ way” of something.

    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
    )
    -1.65
    </h3>
    -1.63
    at
    -1.63
    the
    -1.62
    b
    -1.59
     is
    -1.58
    </h1>
    -1.55
     to
    -1.49
     heridos
    -1.48
    on
    -1.48
    POSITIVE LOGITS
    ☫
    2.17
     lema
    2.03
    其妙
    1.95
    izielle
    1.85
    beitung
    1.77
     Hemos
    1.72
    campista
    1.69
    瘓
    1.66
    ądze
    1.66
    assertj
    1.66
    Activations Density 0.004%

    No Known Activations