© 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. 84464
    Prev
    Next
    INDEX
    Explanations

    medical/technical keywords

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

    formatting tokens and code elements in technical documentation, especially in Asian languages.

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

    The neuron chiefly activates on numeric tokens (i.e. standalone numbers).

    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
    '
    -1.80
    Ŭ
    -1.57
     tersenyum
    -1.55
    獁
    -1.55
     МЕ
    -1.55
     sensibles
    -1.54
    …»
    -1.53
     thuis
    -1.52
     tertawa
    -1.51
     ò
    -1.50
    POSITIVE LOGITS
     others
    2.06
     renowned
    1.71
     name
    1.66
    bigli
    1.66
     więks
    1.64
     same
    1.63
     overarching
    1.63
     dzisiaj
    1.63
    vogue
    1.60
    abriel
    1.59
    Activations Density 0.001%

    No Known Activations