© 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. 10-GEMMASCOPE-RES-131K
    7. 112013
    Prev
    Next
    INDEX
    Explanations

    states, roles, or past actions

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

    The neuron strongly activates on copular “is/are/was” (and similar linking verbs) in definitional or descriptive statements (e.g. “X is Y”).

    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
     such
    -1.54
     some
    -1.33
     their
    -1.13
     it
    -1.13
     a
    -1.09
    大部分
    -1.07
    一部分
    -1.05
     damaligen
    -1.05
     غزة
    -0.97
     those
    -0.97
    POSITIVE LOGITS
     again
    1.34
    Its
    1.31
     beber
    1.23
     nonetheless
    1.16
    times
    1.16
    виси
    1.15
    !».
    1.15
     censo
    1.11
    !
    1.10
    财政
    1.10
    Activations Density 0.070%

    No Known Activations