© 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. 10-GEMMASCOPE-RES-131K
    7. 85383
    Prev
    Next
    INDEX
    Explanations

    sometimes or not always

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

    The neuron responds to adverbial and evaluative modifiers (e.g. words expressing frequency or intensity such as “always,” “sometimes,” “acceptable,” “challenging,” 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_10/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
    Musique
    -1.29
    années
    -1.21
    Geology
    -1.20
    Famille
    -1.16
     phoned
    -1.15
    escence
    -1.15
    gruppo
    -1.11
    ются
    -1.10
    Tongue
    -1.09
    슌
    -1.09
    POSITIVE LOGITS
    even
    1.22
     coisas
    1.20
     your
    1.16
     quando
    1.14
     about
    1.14
     debido
    1.09
     estos
    1.07
    flora
    1.06
     people
    1.06
     hará
    1.05
    Activations Density 0.088%

    No Known Activations