© 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. 21936
    Prev
    Next
    INDEX
    Explanations

    stuck, trapped, entangled

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

    The neuron activates on words describing physical entrapment or obstruction (e.g., trapped, entrapped, impacted, caught, entangled).

    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
    .《
    -1.42
     ==========
    -1.40
     -------------
    -1.38
     nahi
    -1.38
     ã
    -1.35
     nke
    -1.32
     to
    -1.24
     feront
    -1.22
    🔍
    -1.21
     kela
    -1.20
    POSITIVE LOGITS
    ();
    
    
    1.49
     casi
    1.28
     mayhem
    1.26
     entran
    1.26
    何を
    1.19
    複数の
    1.16
     stubborn
    1.16
     must
    1.16
    udert
    1.16
     him
    1.15
    Activations Density 0.087%

    No Known Activations