© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 1750
    Prev
    Next
    INDEX
    Explanations

    resonating, potentiated, patently, crucially, weirder, doled, suboptimal, stomped, phrased, dehumanizing

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

    The neuron fires strongly at the beginnings of longer content words (i.e. the initial character sequences of multi‐syllable words).

    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
    (/^
    -1.08
     differing
    -1.05
     shouldn
    -1.05
    CCIÓN
    -1.05
     feel
    -1.04
    onomies
    -1.03
    があります
    -1.00
    という
    -0.99
     which
    -0.97
    zymes
    -0.97
    POSITIVE LOGITS
     джере
    1.04
    tamount
    1.03
    marily
    1.00
     empêcher
    0.98
    उ
    0.96
    edonia
    0.95
    phur
    0.94
     balm
    0.94
    mbito
    0.94
     indicative
    0.94
    Activations Density 0.110%

    No Known Activations