© 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. 66374
    Prev
    Next
    INDEX
    Explanations

    whitespace and punctuation following

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

    This neuron detects regex character‐class tokens and metacharacters for whitespace and other special symbols (e.g. “\s”, “.”, “_”, “-” inside […]).

    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
    1
    -1.15
     欣赏
    -1.13
     kaynağından
    -1.13
    琑
    -1.12
     Tento
    -1.11
     Če
    -1.09
    </h2>
    -1.08
     EACH
    -1.05
     illet
    -1.04
     danken
    -1.04
    POSITIVE LOGITS
     or
    1.26
    or
    1.20
     yako
    1.17
    jores
    1.17
     between
    1.17
    CELONA
    1.16
     giapponese
    1.14
     giapp
    1.10
     stumbled
    1.08
    衒
    1.08
    Activations Density 0.014%

    No Known Activations