© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34059
    Prev
    Next
    INDEX
    Explanations

    Q: or Eureka!

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

    The neuron fires on bracketed citation or reference markers (e.g., “[ @… ]”) in academic‐style texts.

    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
     maquillage
    -1.13
    疯狂
    -1.06
     Wikipédia
    -1.04
     wciąż
    -0.99
    ähren
    -0.99
    Nazwa
    -0.99
     anniversaire
    -0.98
     vernis
    -0.98
     teatral
    -0.98
     média
    -0.97
    POSITIVE LOGITS
     These
    0.95
    which
    0.93
    XOXO
    0.86
    cciale
    0.86
     religión
    0.86
     and
    0.86
    ora
    0.85
     which
    0.85
     INCLUDES
    0.84
    最新的
    0.83
    Activations Density 0.000%

    No Known Activations