© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 68739
    Prev
    Next
    INDEX
    Explanations

    respectively, lists of numbers

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

    The neuron consistently activates on numeric tokens (especially multi‐digit numbers), i.e. it’s looking for 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
    üğü
    -1.30
    EVEN
    -1.22
     花朵
    -1.18
    脔
    -1.11
    嫆
    -1.09
     kara
    -1.08
     ſal
    -1.06
    diel
    -1.05
     Tá
    -1.04
     ΤΟ
    -1.04
    POSITIVE LOGITS
     respectively
    1.05
     its
    1.02
    }
    
    1.02
     帶
    1.02
     allows
    1.02
     When
    1.01
     when
    1.00
     makes
    0.99
     Before
    0.99
     removed
    0.98
    Activations Density 0.022%

    No Known Activations