© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25197
    Prev
    Next
    INDEX
    Explanations

    too much or too little

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

    The neuron detects English noun-of-gerund constructions—idiomatic phrases like “the habit of thinking” or “the mistake of putting” where a noun is followed by “of” plus a verb-ing.

    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
    さすがに
    -1.15
    Idio
    -1.14
    ooker
    -1.05
    Saman
    -1.03
    いれば
    -1.02
    Vendredi
    -1.02
    Tors
    -1.02
    Samo
    -1.02
    Svet
    -0.99
    玫
    -0.99
    POSITIVE LOGITS
     too
    1.40
     only
    1.34
     فقط
    1.32
     instead
    1.31
     इसे
    1.28
     invece
    1.23
     вместо
    1.21
     слишком
    1.19
    只
    1.17
    低
    1.15
    Activations Density 0.058%

    No Known Activations