© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 74961
    Prev
    Next
    INDEX
    Explanations

    close or terminate code actions

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

    The neuron fires on imperative code tokens that invoke termination or destruction actions (e.g. close, terminate, delete, flush, remove).

    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
    蟥
    -1.39
    
    -1.36
    ྵ
    -1.34
     klasa
    -1.31
    archiwizowane
    -1.30
    ዣ
    -1.29
    ⚌
    -1.26
    裥
    -1.23
    鈇
    -1.22
    푉
    -1.22
    POSITIVE LOGITS
    It
    1.40
    1.33
    (
    1.27
    Don
    1.18
    d
    1.16
    In
    1.15
    m
    1.10
    2
    1.09
     =
    1.09
     bądź
    1.07
    Activations Density 0.000%
    No Known Activations