© 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. 93530
    Prev
    Next
    INDEX
    Explanations

    spoilt, spoiled, spoil, spoiling

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

    This neuron fires on non-content function words—especially subject pronouns (“you,” “we”), auxiliaries and forms of “to be” (“are,” “be,” “will,” “has,” etc.) that mark sentence structure rather than specific lexical content.

    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
    vanju
    -0.79
    шити
    -0.76
    truct
    -0.76
    biotics
    -0.76
    стеро
    -0.76
    pha
    -0.76
    ynes
    -0.76
    ersonal
    -0.76
     بير
    -0.76
    ocular
    -0.75
    POSITIVE LOGITS
     spoilt
    4.16
     spoiled
    4.06
     spoil
    2.61
     spoils
    2.27
    spo
    2.20
     spoiling
    2.14
     riches
    1.81
     Spo
    1.80
    Spo
    1.74
     choices
    1.73
    Activations Density 0.027%

    No Known Activations