© 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. 35656
    Prev
    Next
    INDEX
    Explanations

    frequency of occurrence

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

    The neuron activates whenever it sees numeric tokens or explicit numbers (counts, dates, measurements) in the 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
     još
    -0.84
    PageRoute
    -0.82
     żołnier
    -0.81
     all
    -0.81
    Dichloro
    -0.79
     wojsk
    -0.77
     vš
    -0.76
    많
    -0.76
     לעשות
    -0.76
    허
    -0.75
    POSITIVE LOGITS
     usually
    1.73
     vždy
    1.60
     each
    1.59
     always
    1.59
     often
    1.39
    毎回
    1.37
     всегда
    1.37
     sometimes
    1.31
    usually
    1.31
    Usually
    1.29
    Activations Density 0.061%

    No Known Activations