© 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. 33628
    Prev
    Next
    INDEX
    Explanations

    Imran Khan, official stressed

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

    The neuron strongly activates on capitalized words (especially proper names and organization acronyms) and numeric tokens, i.e. named‐entity and number markers.

    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
     by
    -1.26
    ׇ
    -1.10
     such
    -1.05
     måske
    -1.05
     our
    -1.03
    太多的
    -1.03
    ברית
    -1.02
     their
    -1.02
     Ortsteil
    -1.01
     schreef
    -0.99
    POSITIVE LOGITS
    ּוֹ
    1.65
    ַּ
    1.44
    ִּ
    1.38
    说法
    1.34
    Caracteristici
    1.30
     aforementioned
    1.24
     esportes
    1.23
    ְּ
    1.20
     červen
    1.19
     exercise
    1.18
    Activations Density 0.004%

    No Known Activations