© 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. 22-GEMMASCOPE-RES-131K
    7. 26428
    Prev
    Next
    INDEX
    Explanations

    quantified constituents structure residues

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

    scientific notation and numerical data in research papers.

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

    The neuron primarily detects numerical tokens (e.g. sample sizes, counts, measurements) in scientific text.

    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
    皈
    -1.38
     is
    -1.37
     such
    -1.28
     you
    -1.17
    匾
    -1.14
     {
    -1.09
     How
    -1.05
    工作
    -1.04
    ්
    -1.04
     what
    -1.03
    POSITIVE LOGITS
     overall
    1.27
     všech
    1.25
     وكانت
    1.22
    ținut
    1.22
     bandoulière
    1.22
    髋
    1.19
     eletr
    1.18
     وكان
    1.18
    ֶּ
    1.18
     was
    1.16
    Activations Density 0.012%

    No Known Activations