© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 71581
    Prev
    Next
    INDEX
    Explanations

    recent, latest, highest, closest

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

    The neuron detects superlative degree markers (e.g. “naj‐” in Slavic, “最” in East Asian, “aller-”/“höchste” in German) indicating “most,” “highest,” “recent,” etc.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     to
    -2.55
     of
    -2.41
    狴
    -2.25
     But
    -2.14
    coke
    -2.13
    !«
    -2.11
    )–
    -2.09
     respecting
    -2.03
     diğer
    -2.02
     nimmer
    -1.98
    POSITIVE LOGITS
    $
    2.86
    Almost
    2.77
    Some
    2.75
    相変わらず
    2.75
    Several
    2.67
     perpé
    2.66
    あっという
    2.61
    most
    2.55
    Only
    2.48
    first
    2.44
    Activations Density 0.001%

    No Known Activations