© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 39632
    Prev
    Next
    INDEX
    Explanations

    why is

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

    The neuron specially fires on the copular verb “is,” especially when it appears in questions or headings.

    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
    1
    -2.08
    𔘓
    -2.03
    With
    -1.90
    0
    -1.84
    with
    -1.84
    寕
    -1.84
    Its
    -1.72
    </td>
    -1.71
    い
    -1.71
    crizione
    -1.67
    POSITIVE LOGITS
     it
    3.34
     the
    1.95
     your
    1.74
     this
    1.63
     involunt
    1.57
    ான்
    1.56
    
    1.55
     akut
    1.53
     Which
    1.52
     morfo
    1.52
    Activations Density 0.023%

    No Known Activations