© 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. 67960
    Prev
    Next
    INDEX
    Explanations

    .id, .n, .rg, .my

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

    The neuron detects object or structure member‐access expressions (i.e. tokens where a dot or arrow operator precedes a property or field name).

    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
     lachen
    -0.88
     terdekat
    -0.84
     ilustrasi
    -0.83
     mädchen
    -0.83
    arita
    -0.83
    ด้วย
    -0.82
    aner
    -0.80
    <bos>
    -0.80
    кисло
    -0.79
     împreună
    -0.79
    POSITIVE LOGITS
    音が
    1.16
    hasOwnProperty
    1.05
    ização
    1.03
    線が
    0.96
     предусмотрен
    0.95
     tay
    0.95
    juelas
    0.94
    慳
    0.93
     Tests
    0.93
     impacts
    0.92
    Activations Density 0.009%

    No Known Activations