© 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. 104643
    Prev
    Next
    INDEX
    Explanations

    algorithms and technical terms

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

    function names or technical terms in programming and scientific contexts.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron responds most strongly to unusually long or “multi‐part” tokens (e.g. long function names or lengthy words split across sub‐tokens).

    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
     giebt
    -1.45
     two
    -1.34
     between
    -1.34
     still
    -1.31
     too
    -1.30
     lorsqu
    -1.29
     refiri
    -1.28
     realizado
    -1.26
     ninety
    -1.26
     often
    -1.25
    POSITIVE LOGITS
    5
    1.27
    Cout
    1.22
    位的
    1.19
    ่ง
    1.19
    Chit
    1.17
    esseur
    1.16
    힣
    1.16
     middag
    1.16
    Tamil
    1.16
     toro
    1.16
    Activations Density 0.123%

    No Known Activations