© 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. 34-GEMMASCOPE-RES-131K
    7. 49347
    Prev
    Next
    INDEX
    Explanations

    really hugely amazing

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

    The neuron is picking up on unusual, nonstandard tokens—things like hyphen‐ or soft-hyphen-split words, profanity with embedded symbols, odd numeric/byte sequences, and other out-of-vocabulary or specially formatted character runs.

    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
     aéreo
    -1.54
     untuk
    -1.52
     gekomen
    -1.48
     برای
    -1.48
     supremely
    -1.48
    大家好
    -1.47
    をします
    -1.46
    ֢
    -1.45
     аккура
    -1.45
    はコチラ
    -1.43
    POSITIVE LOGITS
     REALLY
    1.80
     hugely
    1.64
     unbelievably
    1.54
     AMAZING
    1.51
     ridiculously
    1.49
     myriad
    1.46
     wondrous
    1.41
     broader
    1.39
     amazingly
    1.38
     insanely
    1.38
    Activations Density 0.127%

    No Known Activations