© 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. 15959
    Prev
    Next
    INDEX
    Explanations

    dates with numbers and delimiters

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

    The neuron fires on capitalized identifiers or object names in code (e.g. “tree” in JSON nodes or “Template” in VBA).

    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
    ℅
    -1.09
    ту
    -1.07
    "\
    -1.06
    不过
    -1.03
    不過
    -0.96
     tredje
    -0.96
     arkite
    -0.94
    =$((
    -0.94
     العربية
    -0.92
    Anyways
    -0.92
    POSITIVE LOGITS
    坳
    0.91
     faszin
    0.91
    ෛ
    0.88
     dokonce
    0.88
     different
    0.87
     sogar
    0.85
     even
    0.85
     at
    0.85
    了下去
    0.85
     hampered
    0.84
    Activations Density 0.010%

    No Known Activations