© 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. 34-GEMMASCOPE-RES-131K
    4. 90265
    Prev
    Next
    INDEX
    Explanations

    military siege or blockade

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

    The neuron strongly activates on occurrences of the word “siege” (especially in titles like “Siege of X”), i.e. mentions of siege events.

    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
     expanse
    -0.83
     patrolling
    -0.75
    Scrub
    -0.75
     Madeira
    -0.73
     matchups
    -0.71
    arita
    -0.70
     Pasteur
    -0.68
     accum
    -0.68
     distorted
    -0.68
     scrambling
    -0.68
    POSITIVE LOGITS
     siege
    3.33
     Siege
    2.53
     besieged
    2.30
     оса
    2.14
    Siege
    2.11
     besie
    1.97
     sie
    1.78
    sie
    1.55
    围
    1.50
    圍
    1.45
    Activations Density 0.031%

    No Known Activations