© 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. 128949
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly responds to long, multi‐syllabic content words—especially technical or academic nouns.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    technical terms and descriptions

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    <bos>
    -1.34
    才可以
    -1.22
    でも
    -1.15
    や
    -1.13
    が
    -1.09
    も
    -1.08
    は
    -1.05
    其
    -1.03
    Its
    -0.99
     remarkably
    -0.99
    POSITIVE LOGITS
     what
    1.16
     one
    1.08
     at
    1.05
     as
    1.05
     this
    1.02
     it
    1.02
    >`
    0.97
    .”
    0.96
    quests
    0.95
     provide
    0.92
    Activations Density 0.091%

    No Known Activations