© 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. 1992
    Prev
    Next
    INDEX
    Explanations

    per unit

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

    This neuron detects numerical rate expressions of the form “X per Y,” i.e. quantities measured per unit.

    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
    ENSATION
    0.58
     Señ
    0.58
    と言
    0.54
    olini
    0.53
    Shang
    0.52
    Alien
    0.52
    Saund
    0.52
    琊
    0.52
    そうな
    0.52
    ஷ்ய
    0.51
    POSITIVE LOGITS
     per
    1.38
     annum
    1.34
     each
    1.09
     apiece
    1.09
     EACH
    0.97
     capita
    0.96
     mỗi
    0.94
     каждая
    0.94
     каждой
    0.93
     каж
    0.92
    Activations Density 0.287%

    No Known Activations