© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 68745
    Prev
    Next
    INDEX
    Explanations

    actions and states

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

    The neuron selectively activates on words that describe spatial relationships—especially depth ordering and occlusion (e.g. closer/further, block/occlude, depth, view, representation).

    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
     confidently
    -0.95
    ~
    -0.94
     sezione
    -0.93
     (!__
    -0.91
    órmulas
    -0.87
     including
    -0.85
    튼
    -0.85
     관계
    -0.84
     обеспечить
    -0.84
    ~\
    -0.83
    POSITIVE LOGITS
     yani
    1.43
     ("
    1.31
     (“
    1.31
    (“
    1.29
    (=
    1.23
     (‘
    1.22
     aka
    1.20
     cuchara
    1.20
     hatta
    1.13
    aka
    1.11
    Activations Density 0.150%

    No Known Activations