© 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. 39870
    Prev
    Next
    INDEX
    Explanations

    code comment end, question start, abbreviation start

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

    The neuron detects long runs of repeated punctuation characters (e.g. “—”, “/”, “*”) used as section or comment separators.

    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
    なんですね
    -1.14
     bertambah
    -1.09
    Reak
    -1.09
     venturing
    -1.03
    Often
    -1.02
    だそうです
    -1.00
     аккура
    -1.00
    にほんブログ村
    -0.99
    adă
    -0.99
    خول
    -0.98
    POSITIVE LOGITS
     dando
    1.20
    approximately
    1.17
    ha
    1.14
    an
    1.13
    does
    1.13
     immense
    1.11
    jin
    1.10
    considered
    1.09
    hatsu
    1.09
    </em>
    1.09
    Activations Density 0.007%

    No Known Activations