© 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. 26798
    Prev
    Next
    INDEX
    Explanations

    most followed by adjectives

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

    The neuron fires on superlative phrases built with “most” (e.g. “the most seasoned,” “most happiest,” “most pretentious”), highlighting “most” plus its qualifying words.

    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
    และ
    -1.30
    particularly
    -1.14
     և
    -1.11
     instead
    -1.09
     DIFFERENT
    -1.07
     クルー
    -1.07
     唯美
    -1.04
     VARIOUS
    -1.02
     rather
    -1.02
    ⩥
    -1.02
    POSITIVE LOGITS
     of
    1.63
     can
    1.47
    Most
    1.42
     самых
    1.40
    Very
    1.30
    最
    1.27
     most
    1.23
     самые
    1.21
     самы
    1.20
    Through
    1.14
    Activations Density 0.029%

    No Known Activations