© 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. 5662
    Prev
    Next
    INDEX
    Explanations

    vs. and v. in case names

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

    The neuron is strongly activated by legal reference tokens—especially case‐citation numbers, section numbers, and related numeric/legal citation markers.

    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
    └──
    0.59
    ","/
    0.58
     dépl
    0.55
     undergraduate
    0.55
     psyched
    0.54
     displ
    0.54
    utako
    0.53
    rés
    0.53
     ché
    0.53
     convexes
    0.52
    POSITIVE LOGITS
     v
    0.82
     vs
    0.79
     Estate
    0.66
     dicta
    0.66
     Twp
    0.66
     ppl
    0.66
     dictum
    0.64
     Assn
    0.63
     cf
    0.63
     Comune
    0.61
    Activations Density 0.139%

    No Known Activations