© 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. 65678
    Prev
    Next
    INDEX
    Explanations

    serial, parents, newer, this, whilst, because

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

    The neuron fires on single capitalized words—i.e. proper names or other standalone capitalized tokens.

    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
     appétit
    -1.20
     великолеп
    -1.10
     useRouter
    -1.05
     statunitense
    -1.05
     作用
    -1.02
     soggiorno
    -1.02
    欣賞
    -1.00
    場所は
    -1.00
     sürd
    -0.99
     officio
    -0.98
    POSITIVE LOGITS
     is
    1.51
     mivel
    1.11
     النبات
    1.07
    URLEncoder
    1.07
     eher
    1.02
     '
    0.99
     gute
    0.97
     filho
    0.97
     agile
    0.97
    ComponentScan
    0.96
    Activations Density 0.024%

    No Known Activations