© 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. 8053
    Prev
    Next
    INDEX
    Explanations

    word parts forming words

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

    The neuron fires strongly on rare or unusually spelled content words (i.e. low‐frequency, information‐rich nouns and adjectives).

    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
     ܠ
    -1.56
    勿論
    -1.55
     your
    -1.52
    シャレ
    -1.46
    関連リンク
    -1.41
    割と
    -1.41
    かもしれませんが
    -1.40
     ܕ
    -1.39
    そも
    -1.37
    ヨタ
    -1.37
    POSITIVE LOGITS
     
    1.73
     they
    1.63
    𝐋
    1.44
    他们
    1.38
    ４
    1.38
    .”
    1.35
    หลัง
    1.34
    !
    1.33
    ３
    1.30
    ２
    1.27
    Activations Density 0.169%

    No Known Activations