© 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. 36600
    Prev
    Next
    INDEX
    Explanations

    piece goes

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

    This neuron fires on concrete, quantifiable details and technical nouns—especially units, measurements, and structural elements (e.g. “three‐storey,” “gables,” “inches,” “microchip,” “sensor,” “tea cups”).

    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
     umumnya
    -1.17
     in
    -1.12
    涣
    -1.06
    Such
    -1.06
    躋
    -1.04
    Carthy
    -1.02
    咂
    -0.99
    bezpe
    -0.99
    ༼
    -0.98
     or
    -0.98
    POSITIVE LOGITS
     Gesang
    1.22
     morango
    1.20
     razie
    1.18
    ₱
    1.14
    hermosa
    1.14
     perfetto
    1.13
     उसने
    1.13
     wasn
    1.13
     cucchia
    1.11
     soggior
    1.09
    Activations Density 0.021%

    No Known Activations