© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 58573
    Prev
    Next
    INDEX
    Explanations

    LLC abbreviations

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

    The neuron strongly responds to isolated technical identifiers and acronyms (e.g. package names like “gms,” uppercase initialisms like “LJ,” “LLC,” short tokens with diacritics or uncommon letter combinations).

    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
    （※
    -1.53
     emphasizes
    -1.51
    (*´
    -1.48
     for
    -1.45
     emphasized
    -1.44
     Tieren
    -1.44
     stratég
    -1.43
     importantly
    -1.42
    </h4>
    -1.42
     olsa
    -1.41
    POSITIVE LOGITS
    ING
    1.73
    lies
    1.58
    not
    1.57
    分からない
    1.50
    ant
    1.48
    ments
    1.47
    ment
    1.45
     passando
    1.44
     ……
    1.42
    now
    1.41
    Activations Density 0.001%

    No Known Activations