© 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. 70392
    Prev
    Next
    INDEX
    Explanations

    medical, technical, or programming terms

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

    words and phrases that appear in website domains, URLs, and digital content references.

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

    This neuron responds to unusual or low-frequency subword tokens (i.e. rare or “odd” BPE pieces).

    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
    isome
    -1.41
     by
    -1.41
    itively
    -1.32
     alongside
    -1.30
    ѐ
    -1.20
     into
    -1.20
    wigs
    -1.19
    tifiable
    -1.16
    spaces
    -1.16
     hadapan
    -1.16
    POSITIVE LOGITS
     моей
    1.51
     mostaza
    1.47
     posky
    1.43
     veľa
    1.41
     henkilö
    1.40
     désert
    1.37
     많
    1.33
    鋳
    1.33
     why
    1.33
     meinem
    1.31
    Activations Density 0.170%

    No Known Activations