© 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. 72383
    Prev
    Next
    INDEX
    Explanations

    geometric shapes and polygons

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

    The neuron fires on adjectives denoting perfectly regular geometric shapes (e.g. “equilateral,” “regular,” as in “regular hexagon”).

    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
     such
    -1.38
     a
    -1.23
     for
    -1.23
     (
    -1.11
    })
    -1.07
    _
    -1.06
     «
    -1.05
     these
    -1.04
    ents
    -1.04
     "
    -1.04
    POSITIVE LOGITS
     vengo
    1.23
    덫
    1.16
    Biographical
    1.08
    ⍝
    1.08
     colorear
    1.07
     observado
    1.07
     ajustado
    1.05
     preparado
    1.05
    脸上
    1.03
    Răspunde
    1.02
    Activations Density 0.002%

    No Known Activations