© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 4502
    Prev
    Next
    INDEX
    Explanations

    action verbs

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

    The neuron identifies references to judges in legal‐style text (e.g. “J.,” “JJ.,” “Circuit Judges”).

    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_24_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
     ayudar
    0.97
    ந்திய
    0.87
    諦
    0.86
    ierend
    0.81
     möglich
    0.80
     ayud
    0.80
     llegará
    0.78
     ayudará
    0.77
    ্ব
    0.77
     centenary
    0.76
    POSITIVE LOGITS
     Một
    0.80
     F
    0.75
    8
    0.75
    9
    0.74
    5
    0.72
     Ў
    0.71
     
    0.71
    chefs
    0.69
    우
    0.69
    찐
    0.69
    Activations Density 0.003%

    No Known Activations