© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 27087
    Prev
    Next
    INDEX
    Explanations

    programming language keywords

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

    The neuron fires strongly on tokens that introduce top‐level code declarations (e.g. keywords like “var,” “func,” “import,” “class,” “package,” “def,” or other constructs that start 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
    p
    -1.48
    d
    -1.41
     it
    -1.38
    ne
    -1.37
    ch
    -1.36
    H
    -1.34
    h
    -1.33
    C
    -1.33
    de
    -1.31
    ents
    -1.31
    POSITIVE LOGITS
    Buku
    1.51
    Makanan
    1.50
    Gambar
    1.47
    bisous
    1.41
    Bahan
    1.40
    Quoi
    1.39
     工房
    1.38
    Fungsi
    1.38
    Spesifikasi
    1.37
    Terima
    1.35
    Activations Density 0.039%

    No Known Activations