© 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. 22-GEMMASCOPE-RES-131K
    7. 72556
    Prev
    Next
    INDEX
    Explanations

    punctuation followed by capitalized words

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

    The neuron is detecting list‐entry headings or titles separated by a hyphen (e.g., “Artist – Title” or similarly formatted list items).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     QUÉ
    -1.41
     éché
    -1.36
     plazos
    -1.33
     groet
    -1.32
     FÜR
    -1.31
     perioden
    -1.31
     wied
    -1.30
    為に
    -1.29
     pengar
    -1.27
     eene
    -1.27
    POSITIVE LOGITS
     what
    1.73
     so
    1.44
     about
    1.41
     as
    1.36
     how
    1.35
     because
    1.35
     on
    1.34
     when
    1.34
     Because
    1.34
     not
    1.30
    Activations Density 0.018%

    No Known Activations