© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34550
    Prev
    Next
    INDEX
    Explanations

    book or publication identifiers

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

    The neuron is primarily detecting numeric metadata tokens—digits and numbers used in things like ISBNs, publication dates, page counts, and other structured numeric fields.

    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
     Webseiten
    -1.37
     said
    -1.13
    經常
    -1.13
     they
    -1.09
    他說
    -1.05
    解釈
    -1.05
     straips
    -1.04
     Dateien
    -1.03
    𝑞
    -1.02
     meistens
    -1.01
    POSITIVE LOGITS
     הספר
    1.28
     book
    1.24
     this
    1.21
    ὴν
    1.16
     ISBN
    1.14
     ouvrage
    1.13
     boken
    1.12
    龛
    1.10
     devoir
    1.09
     éditions
    1.09
    Activations Density 0.047%

    No Known Activations