© 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. 130605
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on numeric tokens—i.e. digits or multi‐digit numbers representing counts, measurements, dates, etc.

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

    but, and, without

    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
    }
    
    
    -1.13
     torebka
    -0.91
    ;->
    -0.91
     ください
    -0.91
     劇場
    -0.90
     przet
    -0.88
    此外
    -0.88
    oussef
    -0.87
     וכן
    -0.86
     उन्ह
    -0.86
    POSITIVE LOGITS
     but
    1.62
    かもしれませんが
    1.11
     BUT
    1.09
     但是
    1.05
     without
    0.96
    但是
    0.93
    BUT
    0.90
    卻
    0.90
     rechange
    0.90
     querem
    0.88
    Activations Density 0.095%

    No Known Activations