© 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. 103396
    Prev
    Next
    INDEX
    Explanations

    talk or speak

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

    The neuron consistently activates on occurrences of the word “talk” (in any case) in the text.

    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
    Adams
    -0.82
    agraph
    -0.78
    Dispose
    -0.77
     FORMA
    -0.73
    Dawn
    -0.71
     doña
    -0.70
     Nichts
    -0.69
    址
    -0.67
    essi
    -0.66
    Pozdrawiam
    -0.66
    POSITIVE LOGITS
     soon
    0.82
    IsTrue
    0.78
    PRET
    0.71
    arko
    0.70
    コの
    0.69
    としている
    0.68
     Is
    0.68
     talk
    0.68
    нинг
    0.68
    mail
    0.67
    Activations Density 0.041%

    No Known Activations