© 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. 84198
    Prev
    Next
    INDEX
    Explanations

    code parsing and references

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

    This neuron fires on tokens containing non-alphanumeric punctuation (e.g. underscores, dots, slashes, hyphens, parentheses or colons) as found in code identifiers, file paths, and similar strings.

    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
     of
    -1.25
    '
    -1.17
    ’
    -1.13
    ..
    -1.11
    i
    -1.08
     for
    -1.08
    lo
    -1.05
    "
    -1.05
    ”
    -1.04
    g
    -1.03
    POSITIVE LOGITS
    continence
    1.24
    Elektron
    1.17
    Lokasi
    1.16
    ができます
    1.13
    Harga
    1.13
    blusa
    1.11
    Zahl
    1.11
    焄
    1.08
    Книга
    1.08
    Филь
    1.08
    Activations Density 0.120%

    No Known Activations