© 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. 10-GEMMASCOPE-RES-131K
    4. 86926
    Prev
    Next
    INDEX
    Explanations

    keep followed by imperative verbs

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

    The neuron fires on imperative “keep”-style calls to action (e.g. “keep shopping,” “keep going,” “keep watching”).

    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
     contigo
    -1.24
     བ
    -1.20
    :+
    -1.17
     myös
    -1.13
    enskap
    -1.13
    šie
    -1.12
    娣
    -1.10
     .............
    -1.09
    ってしまった
    -1.09
    UserGroup
    -1.07
    POSITIVE LOGITS
     there
    1.39
     используя
    1.28
     []:
    1.26
     môi
    1.08
    toto
    1.06
     please
    1.05
     behaup
    1.05
     Ignore
    1.04
    などに
    1.04
     keep
    1.02
    Activations Density 0.024%

    No Known Activations