© 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. 73542
    Prev
    Next
    INDEX
    Explanations

    granting or allowing statements

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

    The neuron fires on reporting or attribution cues (e.g. “said,” quotation marks, “claimed,” “that”), marking reported speech or attributed statements.

    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
    éndose
    -1.07
    对自己
    -1.03
     etek
    -1.03
     cuando
    -1.02
    ॑
    -0.99
    acyjna
    -0.98
     weiße
    -0.98
     adverten
    -0.95
    weisen
    -0.94
     Ausblick
    -0.93
    POSITIVE LOGITS
     allowing
    2.38
     adding
    1.71
     giving
    1.69
     if
    1.65
     putting
    1.63
     granting
    1.63
     bringing
    1.53
     placing
    1.53
     removing
    1.52
     letting
    1.52
    Activations Density 0.082%

    No Known Activations