© 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. 14399
    Prev
    Next
    INDEX
    Explanations

    hit or impacted

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

    The neuron fires on numeric values—especially figures that follow “hit” or “hits,” i.e. statistical or quantitative numbers linked to being “hit.”

    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
    The
    -2.31
    It
    -2.16
    th
    -2.00
    S
    -1.95
     that
    -1.84
    If
    -1.83
    There
    -1.80
    This
    -1.80
    Despite
    -1.73
    When
    -1.66
    POSITIVE LOGITS
    ೢ
    2.22
    を通して
    2.09
    糌
    2.06
    
    2.03
    閦
    1.94
    龅
    1.94
    
    1.91
     vijf
    1.90
    笮
    1.85
    蟜
    1.84
    Activations Density 0.009%

    No Known Activations