© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 91355
    Prev
    Next
    INDEX
    Explanations

    timeline, expansion, classification, chosen

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

    This neuron consistently activates on longer, multi-syllable content words—especially technical or formal nouns—such as "timeline," "expansion," "interactions," or "classification."

    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
    ;
    -3.77
     of
    -3.22
    ).
    -2.53
    。
    -2.52
     was
    -2.47
     ;
    -2.45
    }$.
    -2.42
     in
    -2.34
    );
    -2.33
    ();
    -2.30
    POSITIVE LOGITS
    ,
    3.30
     rafra
    2.50
     refroid
    2.44
     tromper
    2.19
     étan
    2.16
     rafraî
    2.08
     kapak
    2.06
     coû
    2.05
     réve
    2.03
     manquer
    2.02
    Activations Density 0.979%

    No Known Activations