© 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. 79864
    Prev
    Next
    INDEX
    Explanations

    programming constructs implements interfaces

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

    The neuron detects type-declaration tokens in code—i.e. capitalized class, interface, and module names (and their surrounding generic brackets) in definitions and implements/extends clauses.

    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
    —
    -0.99
    L
    -0.98
    รวจ
    -0.98
    j
    -0.97
    I
    -0.96
    hatenablog
    -0.94
     …
    -0.93
    var
    -0.91
     the
    -0.90
    la
    -0.90
    POSITIVE LOGITS
    Tahun
    1.26
     inš
    1.12
    Ukuran
    1.11
    Pendidikan
    1.10
    Setelah
    1.09
    极为
    1.09
     trimestre
    1.07
     ampliación
    1.05
    atkan
    1.04
    Год
    1.03
    Activations Density 0.009%

    No Known Activations