© 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. 22-GEMMASCOPE-RES-131K
    7. 33263
    Prev
    Next
    INDEX
    Explanations

    Thomas Francis/ITIA

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

    The neuron fires on programming‐language tokens (identifiers and keywords) in code snippets—e.g. parameter names in method bindings or attribute names in XAML such as “to,” “TemplateBinding,” and “Thickness.”

    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
    菝
    -1.82
    (!__
    -1.70
    L
    -1.49
    lijks
    -1.45
    ","
    -1.41
    D
    -1.41
    P
    -1.37
    m
    -1.36
    Notae
    -1.35
    可在
    -1.34
    POSITIVE LOGITS
     the
    1.88
    流石
    1.49
    さすがに
    1.41
    佇
    1.41
    とは言
    1.40
     pensez
    1.30
    したら
    1.30
    風景
    1.28
    了你
    1.27
    齎
    1.27
    Activations Density 0.000%

    No Known Activations