© 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. 129991
    Prev
    Next
    INDEX
    Explanations

    This neuron lights up on specialized multi‐syllable mathematical terminology and formal logical keywords (e.g. “exists,” “every,” “omorphically,” “embeddable,” “complemented”).

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

    embeddable in, retraction of, expressed as

    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
     !!!!!
    -1.09
     hellen
    -0.97
     jezt
    -0.96
    퀼
    -0.95
     correndo
    -0.95
     herren
    -0.95
     myster
    -0.95
     Lorsque
    -0.94
     patrio
    -0.93
     kritis
    -0.93
    POSITIVE LOGITS
     with
    1.56
     so
    1.36
     without
    1.30
     via
    1.19
     can
    1.16
     by
    1.16
     using
    1.15
     but
    1.08
     within
    1.07
     under
    1.03
    Activations Density 0.077%

    No Known Activations