© 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. 129463
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on uncommon, domain-specific keywords—i.e. technical jargon or named concepts across code, scientific, and specialized prose.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    numbers and quantities

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     wsp
    -0.99
     naar
    -0.99
     kontakta
    -0.98
     åren
    -0.96
     fortæller
    -0.95
     kasarigan
    -0.94
     siger
    -0.92
    赣
    -0.91
     januari
    -0.90
     Cong
    -0.90
    POSITIVE LOGITS
     nantinya
    1.21
     both
    1.18
     two
    1.13
     zowel
    1.09
     sowohl
    1.05
     både
    1.00
     wollten
    0.95
     Tanto
    0.95
    both
    0.94
     three
    0.92
    Activations Density 0.049%

    No Known Activations