© 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. 26460
    Prev
    Next
    INDEX
    Explanations

    set or follow (proper noun)

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

    This neuron fires on the key topic words that appear in the titles or headings of questions, effectively spotting the main question‐title terms.

    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
     to
    -1.17
     there
    -1.11
     Webseiten
    -1.04
     these
    -1.03
    一個人
    -1.01
     something
    -0.99
     those
    -0.97
     genuine
    -0.94
     the
    -0.92
     this
    -0.92
    POSITIVE LOGITS
     Фото
    1.34
    руб
    1.22
    през
    1.20
     karena
    1.15
     Ähnliche
    1.14
    та
    1.13
     Souhaitez
    1.13
     chaleureux
    1.13
     Besoin
    1.12
     használat
    1.10
    Activations Density 0.507%

    No Known Activations