© 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. 10-GEMMASCOPE-RES-131K
    7. 75066
    Prev
    Next
    INDEX
    Explanations

    initialization or part

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

    The neuron consistently fires on the very first substantive word of each document or section—that is, it detects the initial content token at the start of a paragraph or segment.

    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
    –
    -2.66
    >
    -2.52
    ならでは
    -2.36
     faithfully
    -2.33
    ↵↵
    -2.33
    -“
    -2.28
    on
    -2.25
     خمس
    -2.23
    ᨆ
    -2.20
     a
    -2.19
    POSITIVE LOGITS
     стоковые
    3.13
    pecabe
    3.00
    什么
    2.47
    等等
    2.41
    檍
    2.39
    纔
    2.38
    Kenapa
    2.34
    ﱠ
    2.34
     endforeach
    2.33
    麪
    2.33
    Activations Density 0.000%

    No Known Activations