© 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. 10-GEMMASCOPE-RES-131K
    7. 101207
    Prev
    Next
    INDEX
    Explanations

    objective c

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

    abbreviated or truncated proper nouns and acronyms, particularly organizational names and brand names.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron selectively activates on capitalized proper names and acronyms (e.g. organization or product names).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     that
    -2.13
     who
    -1.88
    ,
    -1.72
    .$,
    -1.59
    </h5>
    -1.57
    Our
    -1.51
    After
    -1.46
     which
    -1.45
    opheles
    -1.41
    ovico
    -1.39
    POSITIVE LOGITS
     Refle
    1.32
    ichteten
    1.31
     trotzdem
    1.30
    Relo
    1.29
    킵
    1.29
     facendo
    1.28
     troupes
    1.28
    SetOutput
    1.27
    ệu
    1.27
    话
    1.27
    Activations Density 0.130%

    No Known Activations