© 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. 71076
    Prev
    Next
    INDEX
    Explanations

    code structure prefixes

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

    This neuron fires on the first real word of a new line or block (e.g. line-initial tokens).

    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
     their
    -1.53
     delgado
    -1.13
     передний
    -1.05
    važ
    -1.03
     its
    -1.03
    zonder
    -1.02
     cleverly
    -1.01
    她
    -1.01
    Źródło
    -0.99
    utusan
    -0.96
    POSITIVE LOGITS
    有一定
    1.13
     также
    1.13
     ポケット
    1.06
    revealed
    1.04
    менту
    1.02
    iscus
    1.00
    arget
    0.99
    Siamo
    0.98
     berbagai
    0.98
    졍
    0.98
    Activations Density 0.000%

    No Known Activations