© 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. 24-GEMMASCOPE-2-RES-16K
    7. 6301
    Prev
    Next
    INDEX
    Explanations

    response completion

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

    The neuron activates strongly on formal legal-justification language (e.g. “zwingende,” “compelling,” “legitimate grounds,” etc.) indicating when an exception or overriding legal interest is being asserted.

    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
    хбет
    0.75
    )}')
    0.67
    必要があります
    0.65
    elten
    0.64
    ORTH
    0.63
    必要がある
    0.63
     orthon
    0.62
     Experiments
    0.60
    陌
    0.60
    inth
    0.59
    POSITIVE LOGITS
    daki
    0.88
    says
    0.85
    nosi
    0.84
    그리고
    0.81
    ncia
    0.80
     tud
    0.79
     mỹ
    0.79
    ᆼ
    0.79
     एफ
    0.79
    VisualStudio
    0.79
    Activations Density 0.001%

    No Known Activations