© 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. 75481
    Prev
    Next
    INDEX
    Explanations

    said followed by name

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

    The neuron is heavily activated by the punctuation, names, and “said” tokens that occur in speech-attribution phrases (e.g. “, Name said”) within news quotes.

    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
     in
    -1.24
     to
    -1.13
     or
    -1.13
     get
    -1.13
     it
    -1.13
     which
    -1.04
     and
    -0.98
     at
    -0.98
     In
    -0.96
     one
    -0.96
    POSITIVE LOGITS
     SAID
    1.28
     gewi
    1.18
     defire
    1.15
     moft
    1.14
     dette
    1.13
     firft
    1.11
     chofe
    1.10
     senseless
    1.09
     menacing
    1.08
     poffe
    1.07
    Activations Density 0.008%

    No Known Activations