© 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. 22-GEMMASCOPE-RES-131K
    7. 73321
    Prev
    Next
    INDEX
    Explanations

    endings or beginnings

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

    This neuron strongly activates on the main noun (often a gerund or abstract noun) that follows a preposition or appears as the key object in a question or heading.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     tyl
    -0.98
    íticas
    -0.91
     and
    -0.85
    texpr
    -0.84
     همان
    -0.84
    idorm
    -0.84
    帮助
    -0.83
    bentar
    -0.81
     Zaire
    -0.81
     dizaine
    -0.81
    POSITIVE LOGITS
    硖
    1.21
     hice
    1.16
     hablado
    1.16
     correcto
    1.04
    翆
    1.02
    ofing
    1.00
     include
    1.00
    卺
    0.97
     yh
    0.96
     toim
    0.95
    Activations Density 0.120%

    No Known Activations