© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 42894
    Prev
    Next
    INDEX
    Explanations

    ending words like ment, ing, or lepsy

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

    The neuron flags relatively long, uncommon content words (i.e. multi-syllabic technical or rare vocabulary) rather than short or function words.

    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
     différentes
    -1.63
     nouvelles
    -1.50
     sélections
    -1.48
    liegt
    -1.46
     PUSTAKA
    -1.41
    茈
    -1.39
    queurs
    -1.38
    chnung
    -1.38
     témoins
    -1.34
     apache
    -1.34
    POSITIVE LOGITS
     any
    2.00
     you
    1.77
     that
    1.76
     most
    1.56
     my
    1.55
     what
    1.54
     usually
    1.53
     no
    1.52
     intend
    1.41
     its
    1.40
    Activations Density 0.131%

    No Known Activations