© 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. 79693
    Prev
    Next
    INDEX
    Explanations

    padding and margin

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

    This neuron activates on numeric literals (numbers and related assignment tokens) in code, i.e. it picks out numeric values used in property definitions.

    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
     optar
    -1.23
     number
    -1.13
     NDEBUG
    -1.11
     mantenido
    -1.11
     emitir
    -1.10
     who
    -1.09
     transferir
    -1.09
     resistir
    -1.08
     Polskiej
    -1.06
    ०
    -1.06
    POSITIVE LOGITS
     margin
    1.13
     comedor
    1.13
     camioneta
    1.09
     margins
    1.09
     regalo
    1.03
     tryck
    1.01
    isierte
    0.98
    usura
    0.97
     Antibodies
    0.96
     framför
    0.95
    Activations Density 0.022%

    No Known Activations