© 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. 22-GEMMASCOPE-RES-131K
    7. 65328
    Prev
    Next
    INDEX
    Explanations

    importance

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

    The neuron responds to emphatic “importance” markers—essentially spotting occurrences of “important” (and its variants) in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    英國
    -1.13
     haría
    -1.12
    KH
    -1.12
    խ
    -1.10
    ucción
    -1.10
     +
    -1.09
     "}
    -1.07
    பா
    -1.07
    宬
    -1.05
     synonymous
    -1.05
    POSITIVE LOGITS
     important
    1.73
     to
    1.45
    重要的
    1.42
    important
    1.33
     важ
    1.28
    重要な
    1.27
    Important
    1.23
     Important
    1.23
     for
    1.20
    which
    1.20
    Activations Density 0.098%

    No Known Activations