© 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. 129652
    Prev
    Next
    INDEX
    Explanations

    The neuron primarily detects numeric tokens (i.e. numbers and numerical literals, often including time‐unit values).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    time and date units

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     歳
    -0.98
     様
    -0.90
    amblea
    -0.89
     different
    -0.88
     wyg
    -0.87
     theor
    -0.86
     Jahrgang
    -0.86
    Ἰ
    -0.85
     other
    -0.85
     bargain
    -0.85
    POSITIVE LOGITS
    Hilsen
    1.05
    dịch
    1.03
     yesterday
    1.00
    UNLOCK
    0.98
    tokenizer
    0.98
    lish
    0.96
     کوچ
    0.96
     lading
    0.94
    臾
    0.92
     viime
    0.91
    Activations Density 0.007%

    No Known Activations