© 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. 127824
    Prev
    Next
    INDEX
    Explanations

    resolve to, of attributes, last month

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

    The neuron responds most strongly to uncommon content words (i.e. high-IDF or rarely seen words).

    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
     it
    -1.69
     susten
    -1.35
     to
    -1.35
    鰾
    -1.34
     paradas
    -1.32
     careful
    -1.30
     available
    -1.29
    );
    -1.29
    ACCIÓN
    -1.29
    もありますが
    -1.26
    POSITIVE LOGITS
    www
    1.30
     ليس
    1.22
    矚
    1.14
     jener
    1.13
     ännu
    1.11
     angebra
    1.11
     như
    1.10
    9
    1.08
    uens
    1.06
     trennen
    1.05
    Activations Density 0.168%

    No Known Activations