© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 92
    Prev
    Next
    INDEX
    Explanations

    investigation

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

    The neuron is keyed to numeric tokens and digit sequences in the text, strongly activating on numbers.

    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
    feo
    -1.00
    charpe
    -0.99
     осозна
    -0.93
     minutter
    -0.93
     cucchiaio
    -0.87
     сър
    -0.86
    낳
    -0.86
    婭
    -0.85
     їх
    -0.85
    pras
    -0.85
    POSITIVE LOGITS
     みた
    0.91
     iss
    0.88
    ené
    0.87
     beschlossen
    0.86
    chrift
    0.85
    itrile
    0.84
     создать
    0.84
    yttää
    0.84
    лений
    0.83
     )}$
    0.82
    Activations Density 0.005%

    No Known Activations