© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 113471
    Prev
    Next
    INDEX
    Explanations

    actually, real, fact

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

    The neuron detects discourse markers that signal correction or contrast, such as “actually” and “in fact.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     if
    -1.38
     of
    -1.35
     meskipun
    -1.33
     生物
    -1.30
     though
    -1.27
     on
    -1.22
     meski
    -1.21
     桥
    -1.20
    ’).
    -1.19
    +};
    -1.17
    POSITIVE LOGITS
     it
    1.50
    was
    1.30
     encantó
    1.23
    ds
    1.16
     there
    1.14
    深渊
    1.12
    ies
    1.09
    āja
    1.09
    which
    1.07
     which
    1.06
    Activations Density 0.019%

    No Known Activations