© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 54202
    Prev
    Next
    INDEX
    Explanations

    math and code tokens

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

    The neuron is looking for numeric literals (tokens that represent 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
     and
    -1.34
     their
    -1.23
     ongles
    -1.02
     pouvoit
    -1.00
    congratulations
    -1.00
     such
    -0.98
     tertawa
    -0.98
     permiten
    -0.98
     ſuch
    -0.98
    terrasse
    -0.98
    POSITIVE LOGITS
    se
    1.14
    lig
    1.09
    М
    1.02
    ですが
    1.02
    le
    1.00
    За
    0.98
    并非
    0.98
    但是在
    0.97
    isse
    0.96
    В
    0.96
    Activations Density 0.006%

    No Known Activations