© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 13489
    Prev
    Next
    INDEX
    Explanations

    press releases, pressures, and forced actions

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

    The neuron activates strongly on occurrences of the word “press” (as in press releases, press conferences, press images, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     방법에
    0.64
     otto
    0.63
    撐
    0.63
     قاس
    0.61
     cruel
    0.60
     STAND
    0.60
     WILSON
    0.60
     oblig
    0.60
     waist
    0.59
     打
    0.58
    POSITIVE LOGITS
    urized
    0.88
    pressed
    0.68
    ures
    0.64
    pres
    0.63
    graduate
    0.63
     กด
    0.63
    press
    0.62
    ured
    0.61
    lishes
    0.61
    sure
    0.60
    Activations Density 0.121%

    No Known Activations