© 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. 71597
    Prev
    Next
    INDEX
    Explanations

    descriptive states or comparisons

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

    The neuron responds strongly to numeric tokens and measurement values (e.g. integers, decimals, percentages) 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
     will
    -1.31
    Formazione
    -1.18
    According
    -1.17
     fondamentali
    -1.15
     or
    -1.13
     during
    -1.13
     according
    -1.11
     through
    -1.11
     ваз
    -1.11
     According
    -1.10
    POSITIVE LOGITS
    quite
    1.24
    begins
    1.22
    Notice
    1.19
    clearly
    1.14
    shown
    1.10
     here
    1.09
    almost
    1.09
    ARON
    1.08
    seem
    1.08
    ceptable
    1.08
    Activations Density 0.017%

    No Known Activations