© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 11229
    Prev
    Next
    INDEX
    Explanations

    safely, lawfully, legitimately

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

    The neuron strongly activates on terms used in formal discussions of state security powers—especially “intelligence,” “agencies,” and related words concerning government defense or surveillance duties.

    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
    意識
    0.54
     готови
    0.54
     propositional
    0.52
    直接
    0.51
    横
    0.50
     Signific
    0.49
    فور
    0.49
    駄
    0.49
    穿越
    0.49
     concatenation
    0.49
    POSITIVE LOGITS
     legitimate
    1.59
     legít
    1.59
     legitimately
    1.56
     harmless
    1.54
     lawful
    1.39
     justifiable
    1.39
     innocent
    1.33
     lawfully
    1.23
     innoc
    1.23
     legitt
    1.23
    Activations Density 0.562%

    No Known Activations