© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 57765
    Prev
    Next
    INDEX
    Explanations

    marshal to json

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

    The neuron strongly activates on code identifiers for serialization or I/O operations—especially “Marshal” (and to a lesser extent “Read” and “Write”)—i.e. function names handling data (de)serialization.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     engravings
    -0.88
    conector
    -0.85
    Grupos
    -0.85
     Sociedade
    -0.85
    できるように
    -0.84
     Shepherds
    -0.84
    Observaciones
    -0.81
     콘텐츠
    -0.81
    strnc
    -0.80
     Sagittarius
    -0.80
    POSITIVE LOGITS
     JSON
    1.02
     into
    1.00
     Lebanon
    0.88
     INTO
    0.88
     también
    0.87
     Bernard
    0.86
     json
    0.84
    rpura
    0.84
     Mars
    0.84
     Council
    0.83
    Activations Density 0.002%

    No Known Activations