© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 36025
    Prev
    Next
    INDEX
    Explanations

    effects and analysis components

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

    The neuron strongly activates on section-heading and subheading words (often title‐case, single words like “Effects,” “Outcomes,” “Notes,” etc.) in technical or scientific documents.

    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
     because
    -0.93
    They
    -0.89
     their
    -0.87
     surmounted
    -0.85
     since
    -0.83
     has
    -0.80
    注目
    -0.80
    pattern
    -0.79
     resili
    -0.79
    antenimiento
    -0.78
    POSITIVE LOGITS
     neler
    1.08
     både
    1.05
    黯
    0.98
     likes
    0.96
    alanta
    0.93
    advantages
    0.92
    bergement
    0.91
     Vorteile
    0.89
     include
    0.88
     such
    0.87
    Activations Density 0.133%

    No Known Activations