© 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. 10-GEMMASCOPE-RES-131K
    7. 44453
    Prev
    Next
    INDEX
    Explanations

    titles and specific phrases

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

    The neuron strongly activates on punctuation-delimited strings—like text in quotes or parentheses—marking titles of songs, albums, films, episodes, and other named creative works.

    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
     经典
    -1.40
    çage
    -1.40
    不断
    -1.35
    煦
    -1.32
    När
    -1.31
     lorsque
    -1.29
     cuando
    -1.28
     potrebné
    -1.28
     gdy
    -1.27
     gefe
    -1.27
    POSITIVE LOGITS
    /
    1.70
    ,
    1.46
    ricanes
    1.41
    :
    1.35
    ского
    1.32
     technologically
    1.29
    ree
    1.28
    ні
    1.22
     or
    1.22
     necesariamente
    1.19
    Activations Density 0.127%

    No Known Activations