© 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. 71659
    Prev
    Next
    INDEX
    Explanations

    only certain conditions

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

    This neuron strongly activates on isolated single uppercase letters used as list or section labels (e.g., “A,” “B,” “C,” etc.).

    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
    مكن
    -0.84
     from
    -0.81
     rescheduled
    -0.80
    યા
    -0.79
    Decoding
    -0.76
     daarna
    -0.73
     Klassen
    -0.72
     ditambah
    -0.72
     サービス
    -0.71
    acamole
    -0.71
    POSITIVE LOGITS
     only
    1.01
     hardly
    0.99
     nul
    0.88
     csak
    0.88
    PHIA
    0.83
     neither
    0.82
    事は
    0.82
     manuelle
    0.82
     asistente
    0.81
     diseñador
    0.79
    Activations Density 0.016%

    No Known Activations