© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130129
    Prev
    Next
    INDEX
    Explanations

    The neuron fires strongly on words that frame or question a political actor’s stance or attitude (e.g. “no equivocation over his attitude to Russia’s actions,” “does the president think that…”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    accept, believed, shooter

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     杯
    -1.02
    usual
    -1.02
    {{-
    -0.97
    
    -0.96
    getIntent
    -0.96
    菀
    -0.94
     melalui
    -0.94
     FIBA
    -0.92
    argument
    -0.92
    usually
    -0.91
    POSITIVE LOGITS
     erwäh
    1.08
    をご覧ください
    1.00
     the
    0.97
     its
    0.96
     recommend
    0.90
     גם
    0.90
     wording
    0.90
    strated
    0.90
     any
    0.89
     Kỳ
    0.89
    Activations Density 0.055%

    No Known Activations