© 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. 12997
    Prev
    Next
    INDEX
    Explanations

    text formatting and visual emphasis

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

    The neuron detects words that describe typographical emphasis or highlighting (e.g. bold, highlighted, underlined, red).

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

    text formatting instructions or descriptions, particularly mentions of highlighting, bolding, or other visual emphasis applied to text.

    oai_token-act-pair · claude-4-5-sonnetTriggered 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
    าส
    -1.03
     Malden
    -0.98
    撃
    -0.93
    뤄
    -0.93
    ฃ
    -0.91
    lul
    -0.85
     Laur
    -0.85
     pach
    -0.85
    🔭
    -0.85
     rather
    -0.84
    POSITIVE LOGITS
     będziesz
    1.04
     job
    1.02
     wiedzieć
    1.00
    kování
    0.98
    ulaires
    0.94
     نشان
    0.94
    seamnă
    0.93
    getField
    0.90
    *{\
    0.90
     indicate
    0.88
    Activations Density 0.063%

    No Known Activations