© 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. 67728
    Prev
    Next
    INDEX
    Explanations

    department names and topics

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

    The neuron activates on document section headings or titles—capitalized multi‐word phrases that label 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_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
     tört
    -1.16
     tomates
    -1.09
    lindung
    -1.08
     tevé
    -1.04
     mondta
    -1.02
     stratég
    -1.02
     bahkan
    -1.02
     cojín
    -1.01
     vermelhas
    -0.98
    zeigt
    -0.98
    POSITIVE LOGITS
    <eos>
    1.38
     future
    1.05
    他和
    1.00
    inerario
    0.98
     actionPerformed
    0.92
    voegd
    0.91
     for
    0.91
    nąć
    0.91
    nak
    0.89
     главный
    0.89
    Activations Density 0.017%

    No Known Activations