© 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. 22-GEMMASCOPE-RES-131K
    7. 69185
    Prev
    Next
    INDEX
    Explanations

    peaks, values, and scales

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

    The neuron detects technical measurement terms and quantitative descriptors in scientific text—especially words like “peak(s),” “split(s),” “shift,” “value(s),” “fold,” and similar.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     heilig
    -1.38
    CÁ
    -1.34
     sicura
    -1.25
     gestire
    -1.20
     then
    -1.20
     arquitecto
    -1.18
    ּוֹ
    -1.18
     envoyer
    -1.16
     Informações
    -1.13
     gonzález
    -1.12
    POSITIVE LOGITS
     only
    1.16
     entanto
    1.16
    цыі
    1.14
    liputi
    1.13
     войск
    1.12
     چون
    1.12
     at
    1.11
    🅡
    1.11
    Pendant
    1.10
     sogar
    1.09
    Activations Density 0.078%

    No Known Activations