© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 130754
    Prev
    Next
    INDEX
    Explanations

    The neuron activates specifically on the preposition “with.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    only with

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     was
    -1.41
    erobic
    -1.29
     sayfası
    -1.23
    YECTO
    -1.23
     would
    -1.23
    話の
    -1.17
     through
    -1.14
     by
    -1.14
    ","\
    -1.13
     sitesinde
    -1.13
    POSITIVE LOGITS
     только
    1.29
     pouze
    1.27
     тільки
    1.27
     केवल
    1.21
     tylko
    1.21
     only
    1.16
    𝚒
    1.16
     فقط
    1.14
     no
    1.13
    만
    1.13
    Activations Density 0.064%

    No Known Activations