© 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. 67708
    Prev
    Next
    INDEX
    Explanations

    intercepting requests or data

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

    The neuron responds to tokens that name programmatic interception or hooking actions (e.g. “intercept,” “hook,” “break,” “log,” “forward,” “changes”).

    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
    endete
    -1.00
    dit
    -0.87
    irled
    -0.86
    lif
    -0.86
    gate
    -0.83
    サー
    -0.83
    ⊥
    -0.81
    logError
    -0.80
     “
    -0.77
    vė
    -0.76
    POSITIVE LOGITS
     every
    1.51
     your
    1.15
     everyday
    1.09
     dipasang
    1.09
     EVERY
    1.06
     Setiap
    1.04
    Every
    1.04
     their
    1.00
     berpeng
    0.98
    every
    0.98
    Activations Density 0.156%

    No Known Activations