© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 93827
    Prev
    Next
    INDEX
    Explanations

    forming groups or shapes

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

    The neuron fires on technical descriptions of selecting elements at regular intervals—especially color‐theory phrases about colors “opposite,” “equally spaced,” or “forming” shapes around the color wheel.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    FFF
    -0.81
    HttpHeaders
    -0.73
    HttpStatusCode
    -0.72
    omé
    -0.71
     contienen
    -0.70
    sume
    -0.69
     detalj
    -0.69
    ,))
    -0.69
    фере
    -0.69
     Kaufmann
    -0.69
    POSITIVE LOGITS
     forming
    5.78
     Forming
    4.63
    forming
    4.59
     formed
    4.38
    形成
    3.98
     образу
    3.92
     membentuk
    3.92
     formando
    3.92
     form
    3.72
     forms
    3.67
    Activations Density 0.324%

    No Known Activations