© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 104800
    Prev
    Next
    INDEX
    Explanations

    radio car brands

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

    The neuron is primarily detecting code‐style identifiers and acronyms (tokens like CR, coll, Ui, input/output_token, present, required, etc., that appear in programming or markup snippets).

    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
    了的
    -1.31
     fapt
    -1.29
     provato
    -1.27
    fahrene
    -1.23
    KER
    -1.23
    </em>
    -1.21
     ܠ
    -1.19
    thee
    -1.16
    ⠐
    -1.16
    mäh
    -1.13
    POSITIVE LOGITS
    Both
    1.71
    ;
    1.53
     Besides
    1.48
    Besides
    1.48
    When
    1.45
    Those
    1.38
     has
    1.36
     both
    1.34
     zowel
    1.34
     each
    1.30
    Activations Density 0.056%

    No Known Activations