© 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. 10-GEMMASCOPE-RES-131K
    7. 39825
    Prev
    Next
    INDEX
    Explanations

    metrics and qualities

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

    The neuron responds to words naming quantitative metrics or measured categories (e.g. percentage, incidence, index, content, market, composition) in statistical or data‐driven contexts.

    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
    wäsche
    -1.48
    我们
    -1.48
     ghiaccio
    -1.36
     constater
    -1.32
     derde
    -1.30
    íu
    -1.29
    posedge
    -1.28
     zijne
    -1.27
     abbiamo
    -1.27
    æn
    -1.25
    POSITIVE LOGITS
     despite
    1.58
     by
    1.57
    そういった
    1.52
     according
    1.47
     when
    1.43
     in
    1.41
     while
    1.38
    彼らの
    1.38
     even
    1.35
     from
    1.32
    Activations Density 0.243%

    No Known Activations