© 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. 10-GEMMASCOPE-RES-131K
    4. 105625
    Prev
    Next
    INDEX
    Explanations

    names and company titles

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

    The neuron detects isolated title-case or all-caps tokens—like headings, labels, acronyms, company or person names, and phone-number digits—that appear in structured contact or profile sections.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ּוֹ
    -1.74
    וּ
    -1.70
    Välislingid
    -1.59
    くださった
    -1.57
    🪛
    -1.52
     tevreden
    -1.49
     untersucht
    -1.48
    ῦ
    -1.43
    ύ
    -1.43
    וּ
    -1.41
    POSITIVE LOGITS
    ֺ
    1.90
     сахара
    1.53
    ׇ
    1.52
     گیاه
    1.49
     än
    1.48
    =
    
    1.48
    ♾
    1.42
     our
    1.42
     ویتامین
    1.39
     yg
    1.38
    Activations Density 0.225%

    No Known Activations