© 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. 73106
    Prev
    Next
    INDEX
    Explanations

    questions and inquiries

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

    The neuron is selectively activated by question cues—especially interrogative words like why, what, how, do, will, is, and related punctuation.

    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
     and
    -1.13
     passos
    -1.05
    awarkan
    -1.02
    licitations
    -1.02
     การ์ตูน
    -1.02
     selber
    -1.00
     cunos
    -1.00
    Bibliote
    -0.99
     "",
    -0.99
     имел
    -0.99
    POSITIVE LOGITS
     we
    1.41
     why
    1.36
    Aiheesta
    1.35
     here
    1.31
    СТЬ
    1.21
     further
    1.16
     quando
    1.16
     children
    1.13
    県道
    1.11
    meldet
    1.10
    Activations Density 0.004%

    No Known Activations