© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 34959
    Prev
    Next
    INDEX
    Explanations

    vice followed by title

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

    The neuron strongly activates on the title word “Vice,” i.e. occurrences of “Vice” in official ranks (e.g. Vice President, Vice Admiral).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     einzigen
    -0.95
    おそらく
    -0.86
    Luckily
    -0.84
    私には
    -0.82
     wichtigste
    -0.82
    فتر
    -0.80
     많은
    -0.80
     verheir
    -0.79
     combinations
    -0.77
     شاید
    -0.76
    POSITIVE LOGITS
     versa
    1.47
     president
    1.12
     Versa
    1.07
    versa
    0.99
     vice
    0.91
    ‪.‬‬
    0.91
     President
    0.91
     ganda
    0.90
    гипет
    0.89
     chairman
    0.88
    Activations Density 0.006%

    No Known Activations