© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 104719
    Prev
    Next
    INDEX
    Explanations

    university undergraduatesunburnacidClostridiumIDABloodyketoprofen

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

    abbreviations and scientific/medical terms in parentheses.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron selectively activates on uppercase abbreviations, acronyms, and other all-caps technical terms.

    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
     livré
    -1.45
     payé
    -1.38
     gouver
    -1.35
     imprimé
    -1.34
     compliqué
    -1.34
     commandé
    -1.32
     expédié
    -1.30
    こない
    -1.28
     relat
    -1.26
     wither
    -1.23
    POSITIVE LOGITS
     when
    1.49
     even
    1.46
     these
    1.42
     as
    1.41
     particularly
    1.28
     those
    1.26
     nine
    1.23
     at
    1.20
     for
    1.20
     one
    1.20
    Activations Density 0.113%

    No Known Activations