© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 25706
    Prev
    Next
    INDEX
    Explanations

    form optimal, for instance, evaluated twice

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

    The neuron activates on numeric tokens (especially measurements or decimal values) in the text.

    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
    もちろん
    -1.09
    実は
    -1.09
    ここで
    -1.07
     kontroli
    -1.06
     normalen
    -1.06
    很多
    -1.03
     heilige
    -1.03
    更多的
    -1.02
    ว่า
    -1.01
    そんな
    -1.01
    POSITIVE LOGITS
     which
    1.71
     and
    1.45
     this
    1.24
     or
    1.23
     it
    1.18
    ých
    1.15
    vább
    1.12
     gezondheid
    1.11
     because
    1.09
    ՞
    1.09
    Activations Density 0.065%

    No Known Activations