© 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. 22-GEMMASCOPE-RES-131K
    7. 64810
    Prev
    Next
    INDEX
    Explanations

    in fact

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

    This neuron fires on common discourse‐level transition or intensifier phrases (e.g. “in fact,” “most of the time,” “or even”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     PLATES
    -0.89
    ētu
    -0.85
    ⢅
    -0.84
    (:,:,
    -0.84
     transitions
    -0.84
    ッテ
    -0.83
     these
    -0.82
     Mère
    -0.81
    čiai
    -0.81
    cX
    -0.81
    POSITIVE LOGITS
     even
    2.98
     даже
    2.66
     حتی
    2.53
     навіть
    2.39
     bahkan
    2.38
    甚至
    2.33
     nawet
    2.14
     вообще
    2.13
    even
    2.08
     thậm
    2.03
    Activations Density 0.058%

    No Known Activations