© 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. 33889
    Prev
    Next
    INDEX
    Explanations

    directly or simply

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

    The neuron fires on formal “rule‐setting” or classification language—tokens like governing, determining, divided, protocols, sense, oldest—i.e. words used to define rules or categories.

    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
    entown
    -1.14
    不仅
    -1.03
     this
    -1.03
     other
    -1.02
    cologist
    -1.02
    ต่างๆ
    -1.02
     various
    -1.01
    '&
    -1.00
    izator
    -1.00
     lainnya
    -0.98
    POSITIVE LOGITS
     непосредственно
    1.32
    直接
    1.30
     просто
    1.29
     general
    1.23
     traditional
    1.23
     directly
    1.19
    そのまま
    1.16
     literal
    1.13
     直接
    1.10
     simplemente
    1.09
    Activations Density 0.130%

    No Known Activations