© 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. 54339
    Prev
    Next
    INDEX
    Explanations

    special characters

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

    The neuron fires on non‐alphanumeric “filler” tokens—long runs of punctuation or whitespace (e.g. comment‐divider asterisks, hashes, indent spaces) rather than words.

    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
     does
    -0.94
     had
    -0.94
     compares
    -0.92
    “……”
    -0.90
     then
    -0.88
     dadurch
    -0.86
     was
    -0.86
    });
    -0.86
    then
    -0.85
     Hälfte
    -0.85
    POSITIVE LOGITS
    ことも
    1.07
     sentiers
    1.06
    épisode
    1.02
     delikat
    0.99
    évaluation
    0.99
    の小
    0.96
    ejercicio
    0.93
     recev
    0.93
    faux
    0.92
     russia
    0.91
    Activations Density 0.025%

    No Known Activations