© 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. 25428
    Prev
    Next
    INDEX
    Explanations

    lisk account:

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find standalone numeric literals.

    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
     nachher
    -1.15
    不明白
    -1.05
     nowych
    -1.04
     muitas
    -1.02
     duradero
    -1.01
     neuem
    -0.98
     będę
    -0.97
     täglichen
    -0.97
     trochę
    -0.95
     genauso
    -0.95
    POSITIVE LOGITS
     fré
    1.05
     of
    1.05
     stadion
    1.05
     bross
    1.02
     orale
    1.02
     romantis
    1.02
     vernis
    1.01
     plafon
    1.01
     soldat
    1.00
     dezelve
    1.00
    Activations Density 0.005%

    No Known Activations