© 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. 67377
    Prev
    Next
    INDEX
    Explanations

    numbers and positive descriptions

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

    The neuron fires on tokens that are part of decimal‐formatted numbers (particularly the decimal point and adjacent digits).

    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
     after
    -1.42
     before
    -1.22
     jaja
    -1.20
     that
    -1.18
     miliar
    -1.17
    Even
    -1.17
    )».
    -1.17
     stoff
    -1.16
     Although
    -1.16
     hadas
    -1.15
    POSITIVE LOGITS
     ridiculously
    1.42
     wonderful
    1.39
     amazingly
    1.38
     prachtige
    1.38
     vastly
    1.37
     amazing
    1.36
    也好
    1.33
     astonishing
    1.30
     wonderfully
    1.29
     myriad
    1.29
    Activations Density 0.023%

    No Known Activations