© 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. 84079
    Prev
    Next
    INDEX
    Explanations

    `yx` or `ax` followed by punctuation

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

    The neuron strongly activates on short tokens ending with “x” (i.e. two‐letter strings of the form ?x).

    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
    t
    -2.84
    棪
    -2.66
     nearing
    -2.64
     Nearly
    -2.61
     That
    -2.55
     Quite
    -2.52
    婓
    -2.52
     Many
    -2.47
     Numerous
    -2.44
     Многие
    -2.44
    POSITIVE LOGITS
     decidedly
    3.14
     unbelievably
    3.05
    archiwizowane
    2.88
     trois
    2.72
     incredibly
    2.69
     três
    2.59
     The
    2.58
     hefty
    2.56
     soooo
    2.50
     delightfully
    2.50
    Activations Density 0.073%

    No Known Activations