© 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. 10-GEMMASCOPE-RES-131K
    7. 21250
    Prev
    Next
    INDEX
    Explanations

    checking for correctness and compliance

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

    The neuron activates most strongly on formal, procedural or instructional terms (e.g. verbs and nouns used in requirements, guidelines, and “how‐to” contexts).

    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
     MAKES
    -1.63
    probably
    -1.52
    trying
    -1.49
    shows
    -1.48
    possibly
    -1.45
     SHOWS
    -1.41
    makes
    -1.35
    definitely
    -1.34
     créations
    -1.34
    eding
    -1.33
    POSITIVE LOGITS
     where
    1.90
     when
    1.79
     adequate
    1.66
     what
    1.55
     your
    1.53
     sufficient
    1.53
     there
    1.50
     adequately
    1.42
     which
    1.41
     before
    1.38
    Activations Density 0.077%

    No Known Activations