© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 32906
    Prev
    Next
    INDEX
    Explanations

    lengths of time

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

    The neuron fires on numeric tokens (digits) denoting quantities, durations, counts, or measurements.

    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
     first
    -1.13
     new
    -1.07
     NOW
    -1.06
    When
    -1.05
     now
    -1.02
     that
    -0.96
     wcześniej
    -0.96
     distinctly
    -0.95
     higher
    -0.92
     lower
    -0.90
    POSITIVE LOGITS
     until
    1.68
    until
    1.41
     duration
    1.41
     longest
    1.37
     jusqu
    1.35
     till
    1.30
     lasted
    1.28
     até
    1.27
     तक
    1.19
    longer
    1.16
    Activations Density 0.065%

    No Known Activations