© 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. 28402
    Prev
    Next
    INDEX
    Explanations

    file extensions, programming keywords

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

    This neuron activates on individual items in explicit enumerations—e.g. file extensions, tag names, or regex alternatives—essentially tokens listed in array or “A|B|C”‐style constructs.

    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
    為に
    -2.14
    づけ
    -2.05
    にします
    -1.85
    の準備
    -1.79
     of
    -1.73
    際に
    -1.69
    になりたい
    -1.67
    様な
    -1.66
    になれる
    -1.60
    で見る
    -1.58
    POSITIVE LOGITS
    そして
    2.30
    などが
    1.84
    などの
    1.81
    などを
    1.76
     etc
    1.75
    や
    1.72
    を
    1.69
    <bos>
    1.67
    など
    1.55
    <0xA3>
    1.52
    Activations Density 0.026%

    No Known Activations