© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 47495
    Prev
    Next
    INDEX
    Explanations

    numbers and identifiers like errors

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

    The neuron looks for structural/document‐layout markers—things like front‐matter keys (layout, title), section headers (###), separators (---), and other metadata or heading tokens.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    7
    -1.41
     Seventh
    -1.40
     seventh
    -1.20
    
    -1.05
     javax
    -1.04
     seven
    -1.01
    ۷
    -1.00
     seventy
    -0.98
    七
    -0.96
    seventh
    -0.96
    POSITIVE LOGITS
    6
    0.87
    Vesti
    0.71
     spese
    0.69
    Nicole
    0.69
     arrog
    0.69
     Cognac
    0.67
     Loom
    0.66
    4
    0.66
     Orc
    0.65
    Vikt
    0.65
    Activations Density 2.011%

    No Known Activations