© 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. 33582
    Prev
    Next
    INDEX
    Explanations

    files, libraries, data, and conditions

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

    This neuron fires on the head word of multi-word or hyphenated technical terms (e.g. the final noun in “directory structures,” “standard library,” “upper respiratory tract,” “remotely-managed workers,” etc.).

    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
     poteva
    -0.93
     seçenek
    -0.87
     the
    -0.87
     when
    -0.86
     seems
    -0.85
    allin
    -0.85
    特的
    -0.84
     because
    -0.84
    --
    -0.84
     ...
    -0.84
    POSITIVE LOGITS
    дъ
    1.14
    </tfoot>
    1.12
    脔
    1.09
    洌
    1.06
    唿
    1.03
     thoſe
    0.99
     коля
    0.99
    ၊
    0.98
     meille
    0.97
    menistan
    0.97
    Activations Density 0.208%

    No Known Activations