© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 129865
    Prev
    Next
    INDEX
    Explanations

    Singleton, react

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

    The neuron fires on code‐style tokens (identifiers, type names, file markers, version numbers, etc.) rather than natural‐language words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     relâche
    -1.74
     schreck
    -1.73
     kertoo
    -1.68
     bailar
    -1.66
    沨
    -1.65
    🪳
    -1.63
    ،
    -1.56
    ᐕ
    -1.48
     знали
    -1.48
    колко
    -1.48
    POSITIVE LOGITS
    at
    1.76
    man
    1.66
    er
    1.64
     emot
    1.62
    ut
    1.62
     hecta
    1.53
    no
    1.52
     faa
    1.52
    men
    1.52
     tdi
    1.50
    Activations Density 0.001%

    No Known Activations