© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 31584
    Prev
    Next
    INDEX
    Explanations

    sentence endings and interjections

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

    The neuron strongly activates on tokens that serve as informal emphasis or discourse markers—e.g. ellipses, exclamation points, parentheses, and similar tone-marking symbols.

    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
     FUCK
    -1.14
     halle
    -1.13
     omg
    -1.07
    Omg
    -1.04
    OMG
    -1.01
    entwick
    -1.00
    FUCK
    -1.00
     exhort
    -0.99
     randon
    -0.98
     HOLY
    -0.98
    POSITIVE LOGITS
     nice
    2.34
     typical
    1.98
     good
    1.89
     Nice
    1.84
     way
    1.74
    typical
    1.56
     classic
    1.55
    Nice
    1.55
     Typical
    1.52
    Typical
    1.48
    Activations Density 0.023%

    No Known Activations