© 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. 109266
    Prev
    Next
    INDEX
    Explanations

    understanding of subjects and their nature

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

    The neuron fires strongly on high‐level conceptual or technical nouns and adjectives—terms like “true,” “market,” “volatile,” “state,” “issue,” “component,” or “cost”—that signal key domain concepts in a discussion.

    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
    ellige
    -0.90
    着一个
    -0.88
     [],
    -0.87
    ۵
    -0.84
     réus
    -0.84
    já
    -0.83
     krijgt
    -0.82
    튀
    -0.82
    jwa
    -0.81
    키
    -0.81
    POSITIVE LOGITS
     true
    1.61
     nature
    1.59
     meaning
    1.54
     significance
    1.46
     situation
    1.41
     magnitude
    1.40
     way
    1.39
     workings
    1.38
    真正的
    1.35
     strengths
    1.34
    Activations Density 0.065%

    No Known Activations