© 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. 128279
    Prev
    Next
    INDEX
    Explanations

    This neuron consistently activates on the running prose of an article’s main body—i.e. the actual narrative text—while ignoring headers, bylines, menus, and other non-story elements.

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

    words following "wondering"

    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
    subaru
    -1.17
     would
    -1.15
    !”
    -1.12
     suceso
    -1.12
     mereka
    -1.09
     they
    -1.09
    招生
    -1.07
     updateTime
    -1.06
    是非常
    -1.05
     unserem
    -1.05
    POSITIVE LOGITS
     another
    1.45
     APPELLANT
    1.23
    dicionado
    1.13
    nadequate
    1.13
    athered
    1.12
     İyi
    1.12
     irgendwann
    1.09
    mogen
    1.05
     hectáreas
    1.03
    Chất
    1.02
    Activations Density 0.048%

    No Known Activations