© 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. 72979
    Prev
    Next
    INDEX
    Explanations

    committee meetings, system resources, or party

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

    The neuron fires most strongly on the head noun in two‐word technical or formal compound terms (e.g. “unified memory,” “Steering Committee,” “set‐top boxes”).

    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
    Herkunft
    -1.17
     discí
    -1.13
    鰻
    -1.12
    共演
    -1.10
     ubezpiec
    -1.09
     "'";
    -1.04
     aporte
    -1.03
     femenina
    -1.00
     desped
    -0.98
     Añade
    -0.98
    POSITIVE LOGITS
    antaranya
    1.07
    |.
    1.05
    leggen
    1.05
    penuhi
    1.05
    joyed
    1.02
    様な
    1.01
    1.01
    0.99
    嗒
    0.99
    と呼ばれる
    0.99
    Activations Density 0.073%

    No Known Activations