© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34882
    Prev
    Next
    INDEX
    Explanations

    multi-lingual identifiers and specific start of words

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

    The neuron detects substrings of garbled alphanumeric data typical of Base64 or other binary-encoded 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
     for
    -1.12
     then
    -1.09
    ulously
    -1.04
     who
    -1.01
    simage
    -0.99
    ritsar
    -0.99
     stratég
    -0.98
     igreja
    -0.96
     creó
    -0.95
     doskona
    -0.95
    POSITIVE LOGITS
    Є
    1.05
    今回は
    1.04
    licopter
    1.02
     seksi
    0.97
     paroi
    0.96
    Longest
    0.95
    ここでは
    0.94
    árqu
    0.93
     обычно
    0.91
     Parece
    0.91
    Activations Density 0.001%

    No Known Activations