© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 64567
    Prev
    Next
    INDEX
    Explanations

    frame object, Rosewood, Kiev

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

    The neuron strongly activates on numeric tokens—i.e. standalone numbers or numeric literals.

    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
     their
    -1.59
     like
    -1.45
     for
    -1.29
     they
    -1.27
     from
    -1.21
     孩子
    -1.21
    inawa
    -1.14
     睡
    -1.14
     when
    -1.11
     mereka
    -1.11
    POSITIVE LOGITS
     devra
    1.40
    まぁ
    1.23
     prestig
    1.23
     tremend
    1.22
    如果
    1.16
    seems
    1.15
    轮回
    1.14
    slightly
    1.14
    几人
    1.13
    Mink
    1.13
    Activations Density 0.000%

    No Known Activations