© 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. 103432
    Prev
    Next
    INDEX
    Explanations

    things, objects, modified, Hyperolius

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

    Japanese words or phrases that appear in text, especially when they're followed by their English translations or explanations.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron detects tokens meaning “thing” or “object” in East Asian languages (e.g. Japanese もの/mono and Chinese 东西).

    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
     paff
    -2.58
     aussit
    -2.56
    龗
    -2.55
    ༵
    -2.50
     ftre
    -2.50
     etui
    -2.50
     hins
    -2.48
    蔀
    -2.48
    櫺
    -2.42
     herren
    -2.39
    POSITIVE LOGITS
    …
    3.02
    </h3>
    2.77
     echten
    2.47
    jenigen
    2.22
     passenden
    2.22
    ,’
    2.20
     geringen
    2.19
     —
    2.19
     leichten
    2.13
    {
    2.11
    Activations Density 0.000%

    No Known Activations