© 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. 39843
    Prev
    Next
    INDEX
    Explanations

    struct definition or instance initialization

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

    The neuron fires strongly on “=” signs and adjacent tokens where variables or constants are being defined or initialized, i.e. it detects assignment/initialization expressions in code.

    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
    ěř
    -1.12
     extérieure
    -1.02
    —
    -1.01
    笑容
    -0.98
     your
    -0.96
     пыта
    -0.95
     Rohit
    -0.94
     zrobić
    -0.94
     delante
    -0.93
    queryInterface
    -0.93
    POSITIVE LOGITS
     mondi
    1.03
    ......”
    1.02
     perna
    0.98
     tait
    0.98
    (^^
    0.97
     alpina
    0.97
    ↓↓↓
    0.97
     forêts
    0.95
    Gente
    0.94
    Other
    0.94
    Activations Density 0.002%

    No Known Activations