© 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. 72384
    Prev
    Next
    INDEX
    Explanations

    affirmations

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

    The neuron primarily detects question marks (i.e. identifies when a sentence is a question).

    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
     fotografi
    -1.01
    entwick
    -0.95
     parfü
    -0.94
    lösung
    -0.93
     gece
    -0.93
    ˄
    -0.92
     parfüm
    -0.91
     campeón
    -0.90
    ”—
    -0.90
    bä
    -0.90
    POSITIVE LOGITS
     yeah
    1.59
     Yes
    1.42
     that
    1.40
     Yeah
    1.40
     yes
    1.38
    Yeah
    1.25
     well
    1.20
    Yes
    1.16
     okay
    1.10
    yes
    1.05
    Activations Density 0.034%

    No Known Activations