© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 130389
    Prev
    Next
    INDEX
    Explanations

    The neuron flags short all-caps (or mixed-case) acronyms and initialisms (e.g. ACPI, ACF, AQP) in technical or scientific contexts.

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

    ac, aq, af, acyl, ACF

    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
    鹮
    -2.83
    ’，
    -2.73
    又不
    -2.56
    丁寧に
    -2.44
    袛
    -2.42
    这也
    -2.41
    のが
    -2.38
     Of
    -2.23
     mobiliario
    -2.23
    ates
    -2.16
    POSITIVE LOGITS
    At
    2.39
     This
    2.36
    垍
    2.22
     Besonder
    2.20
    閦
    2.17
    As
    2.14
    𝟐
    2.11
     devote
    2.11
    鰱
    2.09
    ohjel
    2.08
    Activations Density 0.110%

    No Known Activations