© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 105411
    Prev
    Next
    INDEX
    Explanations

    code comment separators, delimiters

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

    lines of repeated separator characters (dashes, equals signs, asterisks) commonly used in code comments and document formatting.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    This neuron detects long runs of repeated punctuation characters (like sequences of hyphens, asterisks, or slashes) used as separator lines.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     recomendar
    -1.60
     animado
    -1.57
     fantástica
    -1.51
     divertido
    -1.50
    vz
    -1.49
     brilh
    -1.45
     divertida
    -1.42
     encantador
    -1.42
     府
    -1.42
     pesado
    -1.41
    POSITIVE LOGITS
     are
    1.56
    4
    1.38
     podría
    1.30
     moti
    1.28
     we
    1.26
    是
    1.26
     tots
    1.23
    有任何
    1.21
    Если
    1.20
    1
    1.20
    Activations Density 0.035%

    No Known Activations