© 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. 68857
    Prev
    Next
    INDEX
    Explanations

    binary data types and representation

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

    The neuron fires most strongly on code tokens dealing with binary‐data handling—imports of binary/encoding packages, byte- and integer-type names (like uint64/ieee754), and memory-load or unaligned-load operations.

    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
    children
    -1.18
    dä
    -1.16
     todas
    -1.15
    ชื่อ
    -1.15
     enfants
    -1.14
    voked
    -1.10
     recursos
    -1.09
    ecken
    -1.08
     offer
    -1.06
    students
    -1.05
    POSITIVE LOGITS
     from
    1.45
     verksamhet
    1.12
     Alignment
    1.09
    そうに
    1.06
     utbild
    1.06
     företag
    1.01
     ország
    1.00
     Gypsy
    0.99
     behandeln
    0.98
     gyors
    0.96
    Activations Density 0.013%

    No Known Activations